SYNTHESIS, STRUCTURAL CHARACTERIZATION AND SPECTROSCOPIC STUDY OF NANOCRISTALLINE AND MICROCRYSTALLINE MATERIALSmicrocristales

UNIVERSIDAD DE CANTABRIA
Facultad de Ciencias
Departamento de Fı́sica Aplicada
Tesis Doctoral
SYNTHESIS, STRUCTURAL CHARACTERIZATION AND

SPECTROSCOPIC STUDY OF NANOCRYSTALLINE
AND MICROCRYSTALLINE MATERIALS
Rosa Martı́n Rodrı́guez
Santander, Noviembre de 2010

D. Rafael Valiente Barroso, Profesor Titular de Fı́sica Aplicada
INFORMA:
Que el trabajo que se presenta en esta memoria, titulado SÍNTESIS, CARACTERI-

ZACIÓN ESTRUCTURAL Y ESTUDIO ESPECTROSCÓPICO DE MATE-
RIALES NANOCRISTALINOS Y MICROCRISTALINOS, ha sido realizado bajo
su dirección en el Departamento de Fı́sica Aplicada de la Universidad de Cantabria, y
emite su conformidad para que dicha memoria sea presentada y tenga lugar, posterior-
mente, la correspondiente lectura y defensa.
Fdo.: Rafael Valiente Barroso

Table of Contents
Table of Contents i
Acknowledgements v
List of Abbreviations vii
Resumen 1
1 Introduction 5
2 Theory 9
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Interest of nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 Insulating materials. Optical properties . . . . . . . . . . . . . . . . . . . . 11
2.3.1 Doped insulators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.2 Ions in a static crystalline environment . . . . . . . . . . . . . . . . 11
2.3.3 The configurational coordinate diagram . . . . . . . . . . . . . . . . 12
2.3.4 Light absorption and emission processes . . . . . . . . . . . . . . . 14
2.3.5 Non-radiative transitions . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.6 Dimers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Upconversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6 Rare-earth ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.7 Transition metal ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.8 Semiconductor nanostructures . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.9 Interband absorption in semiconductors . . . . . . . . . . . . . . . . . . . . 26
2.9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.9.2 Band edge absorption in direct gap semiconductors . . . . . . . . . 27
2.9.3 Band edge absorption in indirect gap semiconductors . . . . . . . . 28
2.10 Structural phase transitions. Equations of state . . . . . . . . . . . . . . . 29
Bibliography 31
i
ii CONTENTS
3 Synthesis 37
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2 Mechano-chemical processes in a planetary ball mill . . . . . . . . . . . . . 37
3.2.1 Er3+ , Yb3+ co-doped Y2 O3 . . . . . . . . . . . . . . . . . . . . . . . 39
3.2.2 Er3+ , Yb3+ co-doped NaYF4 . . . . . . . . . . . . . . . . . . . . . . 39
3.2.3 Pure and Yb3+ -doped CdS . . . . . . . . . . . . . . . . . . . . . . . 40
3.3 Combustion reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.4 Pechini’s method (sol-gel) . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.5 Nanoparticles coating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.6 Precipitation method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.7 Colloidal semiconductor nanocrystals . . . . . . . . . . . . . . . . . . . . . 45
Bibliography 47
4 Experimental Methods 51
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2 Structural characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2.1 X-ray diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2.2 Transmission electron microscopy . . . . . . . . . . . . . . . . . . . 54
4.3 Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3.1 Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3.2 Luminescence and excitation . . . . . . . . . . . . . . . . . . . . . . 57
4.3.3 Lifetime and time resolved spectroscopy . . . . . . . . . . . . . . . 59
4.3.4 Raman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.4 Temperature dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.5 High pressure measurements . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.5.1 Diamond anvil cells . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.5.2 Pressure calibration and transmitting media . . . . . . . . . . . . . 67
4.6 Spectra correction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.6.1 Wavelength correction . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.6.2 Intensity correction . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Bibliography 70
5 Insulating materials. Luminescent properties 73

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.2 Rare-earth ions doped nanoparticles . . . . . . . . . . . . . . . . . . . . . . 73
5.2.1 Er3+ , Yb3+ co-doped Y2 O3 . . . . . . . . . . . . . . . . . . . . . . . 75
5.2.2 Er3+ , Yb3+ co-doped NaYF4 . . . . . . . . . . . . . . . . . . . . . . 89
5.2.3 Tb3+ or Eu3+ and Yb3+ co-doped Gd3 Ga5 O12 and Y3 Al5 O12 . . . . 97
5.3 Rare-earth and transition-metal ions co-doped systems . . . . . . . . . . . 112
5.3.1 Mn2+ , Yb3+ co-doped LaMgAl11 O19 . . . . . . . . . . . . . . . . . . 113
5.4 Transition-metal ions doped nanoparticles . . . . . . . . . . . . . . . . . . 129
5.4.1 Cr3+ -doped Gd3 Ga5 O12 . . . . . . . . . . . . . . . . . . . . . . . . 130
CONTENTS iii
Bibliography 140
6 Semiconductor nanoparticles. Optical spectroscopy at high pressure 149

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.2 CdS nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.2.1 Synthesis and characterization . . . . . . . . . . . . . . . . . . . . . 150
6.2.2 Optical properties and X-ray diffraction under high pressure . . . . 152
6.2.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
6.3 Colloidal nanocrystalline Zn1−x Cox O . . . . . . . . . . . . . . . . . . . . . 160
6.3.1 Synthesis and characterization . . . . . . . . . . . . . . . . . . . . . 160
6.3.2 Optical properties under high pressure . . . . . . . . . . . . . . . . 163
6.3.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
Bibliography 176
7 Conclusiones 185
List of Publications 189

iv CONTENTS
Acknowledgements
Esta tesis es hoy una realidad gracias a la ayuda de mucha gente, por ello, antes de nada,
me gustarı́a dedicar unas pocas lı́neas de agradecimiento a todas aquellas personas que
tanto en el ámbito cientı́fico como en el personal han contribuido a este trabajo.
Mi director de tesis, Rafael Valiente, por haberme dado la oportunidad y la confianza
necesarias para la realización de este trabajo, por sus buenas ideas, por su ayuda en el
laboratorio y por enseñarme gran parte de lo que he aprendido en este tiempo.
Fernando Rodrı́guez, quién me introdujo en esto de la investigación, por estar siempre
ahı́ cuando se le necesita.
Jesús González, por el entusiasmo que demuestra y transmite a quién le rodea y porque
ha sido una suerte contar con su apoyo en la parte final de esta tesis.
Todos los miembros de los departamentos de Fı́sica Aplicada y CITIMAC, por estar
siempre dispuestos a echar una mano y porque han contribuido a crear un ambiente
realmente agradable. Un recuerdo especial a Fernando Aguado, por su ayuda con los
rayos X bajo presión, Jose I. Espeso, Toño, Luis Echeandı́a, José Manuel, Nieves y Rosa.
Carmen Blanco, Carmen Pesquera y Fernando González del Dpto. de Ingenierı́a
Quı́mica y Quı́mica Inorgánica de la UC por su ayuda en mis inicios en la sı́ntesis quı́mica
de nanopartı́culas de Y2 O3 .
Marco Bettinelli dell’Università di Verona e tutti i membri del suo gruppo, in parti-
colare Adolfo Speghini, Fabio Piccinelli e Erica Viviani per il loro aiuto durante il mio
soggiorno a Verona e per l’interesse dimostrato per il mio lavoro. Nel suo laboratorio sono
stati sviluppati i campioni di YAG, GGG and LMA raccolti in questa tesi.
Daniel R. Gamelin from the University of Washington for all the opportunities he
offered me during my stay in Seattle and all the members of his research group, espe-
cially Mike White, for his help in the preparation and characterization of ZnO colloidal
nanoparticles.
v
El servicio de Microscopı́a Electrónica de Transmisión (SERMET) de la UC por las
imágenes TEM de mis muestras nanocristalinas.
Valentı́n Garcı́a Baonza ası́ como David Santamarı́a e Ignacio López, de la Universidad
Complutense de Madrid por su ayuda con las medidas de difracción de rayos X bajo
presión.
Alfredo Segura y Gloria Almonacid de la Universidad de Valencia por la colaboración
en las medidas de absorción bajo presión en el ZnO.
Joaquı́n Fernández, Rolindes Balda, y Sara Garcı́a-Revilla, de la Universidad del Paı́s
Vasco por su colaboración en las medidas del tiempo de vida del Yb3+ .
Todos los becarios, compañeros de Aplicada y CITIMAC, que saben mejor que nadie lo
que esto supone y que sin duda han contribuido a que estos años estén llenos de buenos y
divertidos recuerdos. Me gustarı́a destacar a Susana, Diego, Braulio, Eva, Marta, Álvaro,
Pablo, Carlos, Cristina, Omar, Marı́a . . . y muchos más teniendo en cuenta los que ya no
están en la UC y los estudiantes extranjeros con los que también hemos compartido este
tiempo.
El resto de mis amigos, los de la infancia, los de la facultad, los nuevos, los que están
lejos . . . por ser también de alguna manera partı́cipes de esto.
Toda mi familia, especialmente Pili que es como una hermana para mı́, y Mónica por
su ayuda con el inglés, ası́ como Carlos y Lines.
Y por último, las personas más importantes de mi vida. Mis padres, porque me lo
han dado todo, me han apoyado siempre, me han enseñado lo realmente importante y
porque lo que soy se lo debo a ellos. Y Guille, por estar siempre a mi lado con una sonrisa
en los buenos momentos y en los que no lo son tanto, y porque todo lo que hago tiene
más sentido porque puedo compartirlo con él. Seguiremos cumpliendo objetivos juntos.
A ellos tres, les dedico este trabajo.
Esta tesis doctoral ha sido elaborada gracias a una beca de investigación dentro del
Programa de Formación de Personal Investigador del Ministerio de Educación y Ciencia
(Ref. BES-2006-13359).
vi
List of Abbreviations
AP Ambient pressure
CA Citric acid
CB Conduction band
CCD Charge-coupled device
CR Cross-relaxation
CW Continuous wave
DAC Diamond anvil cell
DMSO Dimethyl sulfoxide
DMSs Diluted magnetic semiconductors
e-h Electron-hole
ESA Excited-state absorption
ESCO Excited-state crossover
ETU Energy transfer upconversion
G/N Glycine-to-nitrates ratio
GGG Gd3 Ga5 O12
GSA Ground-state absorption
IR Infrared
vii
LA Longitudinal acoustic
LD Laser-diode
LO Longitudinal optic
LMA LaMgAl11 O19
LMnA LaMnAl11 O19
MIR Medium infrared
NIR Near infrared
ODE Octadecene
OPO Optical parametric oscillator
PEG Polyethylene glycol
PMT Photomultiplier
PTFE Polytetrafluoroethylene
QDs Quantum dots
RE Rare earth
RS Rock-salt
RT Room temperature
TA Transverse acoustic
TEM Transmission electron microscopy
TEOS Tetraethyl orthosilicate
TM Transition metal
TO Transverse optic
TOPO Trioctylphosphine oxide
viii
UC Upconversion
UV Ultraviolet
VB Valence band
W Wurtzite
XRD X-ray diffraction
YAG Y3 Al5 O12
ZB Zinc-blende
ZPL Zero phonon line
ix
Resumen
El trabajo resumido en esta tesis se ha llevado a cabo dentro del grupo de Altas Presiones
y Espectroscopia de la Universidad de Cantabria y tiene dos objetivos principales. Por un
lado, se han estudiado las propiedades ópticas de distintos iones de metales de transición
y tierras raras en diversas redes aislantes. Se han realizado medidas experimentales de
luminiscencia, absorción, tiempo de vida y espectroscopia resuelta en tiempo. Concreta-
mente, uno de los objetivos principales ha sido establecer los mecanismos responsables de
la luminiscencia de upconversion (UC) de los materiales estudiados, y los requerimientos
estructurales para aumentar su eficiencia. Para ello se han comparado diversos métodos
de sı́ntesis ası́ como diferentes concentraciones de impurezas. Asimismo, se han investi-
gado transiciones de fase a alta presión en semiconductores con tamaño de partı́cula en el
rango de los nanómetros mediante absorción óptica, espectroscopia Raman y difracción
de rayos X.
Los procesos de UC permiten obtener luz visible de forma eficiente tras la excitación
en el infrarojo (IR). El estudio de materiales que presentan luminiscencia de UC atrae
un gran interés por sus posibles aplicaciones tales como materiales láser, marcadores
biológicos, fósforos o para mejorar la eficiencia de células solares.
La mayorı́a de los estudios de UC involucran combinaciones de iones de tierras raras,

en particular, el Yb3+ es un excelente candidato para inducir UC en otros iones. En
este trabajo se han caracterizado estructural y espectroscópicamente nanopartı́culas de
Y2 O3 : Er3+ , Yb3+ , NaYF4 : Er3+ , Yb3+ , ası́ como nanocristales de Gd3 Ga5 O12 (GGG)
e Y3 Al5 O12 (YAG) co-dopados con Tb3+ -Yb3+ y Eu3+ -Yb3+ preparadas usando diversos
métodos de sı́ntesis. Los procesos de UC en los materiales dopados con iones que poseen
niveles de energı́a resonantes, tales como Er3+ e Yb3+ , se han asignado a mecanismos de
1
2
absorción de estado excitado (GSA/ESA) y transferencia de energı́a (GSA/ETU). Por

el contrario, la luminiscencia de UC en los sistemas con Tb3+ -Yb3+ o Eu3+ -Yb3+ se ha
explicado de acuerdo con el mecanismo de sensitización cooperativa.
Al contrario que las transiciones f-f en las tierras raras, las transiciones d-d de los
metales de transición son mucho más sensibles al entorno, por ejemplo, el Mn2+ puede
emitir desde la región azul hasta la roja del espectro, dependiendo del campo cristalino
generado por la red en la que se encuentre. Por ello, la combinación de iones de lantánidos
y metales de transición extiende las posibilidades de sintonización de la luminiscencia de
UC a través de cambios en las energı́as de los estados involucrados. La emisión de UC en
sistemas mixtos metal de transición-lantánido fue observada por primera vez en CsMnCl3
y RbMnCl3 dopados con Yb3+ sólo a bajas temperaturas. En este trabajo se ha observado
por primera vez luminiscencia verde de UC en el sistema LaMgAl11 O19 (LMA): Mn2+ ,
Yb3+ hasta 650 K. Asimismo se ha llevado a cabo un estudio espectroscópico exhaustivo
de este material analizando la dependencia de sus propiedades de UC con la concentración
de impurezas y la temperatura.
Las técnicas de alta presión pueden modificar la simetrı́a y la fuerza del campo crista-
lino de un material, por ello resultan muy útiles a la hora de obtener información acerca
de las propiedades ópticas de un ión, especialmente en el caso de los metales de tran-
sición. En este trabajo se han investigado las propiedades ópticas de nanopartı́culas de
GGG impurificadas con Cr3+ en función de la temperatura (en el rango 25-300 K) y de la
presión (hasta 20 GPa). Este sistema resulta particularmente interesante debido a que el
Cr3+ se encuentra cerca del punto de cruzamiento de estados excitados 4 T2 -2 E y el estado
desde el que tiene lugar la emisión es una mezcla de ambos estados excitados.
Por último, se han estudiado las propiedades ópticas y estructurales de nanopartı́culas

de CdS y Zn1−x Cox O a alta presión. Las propiedades de los semiconductores nanocris-
talinos pueden ser muy distintas de las del material masivo (bulk ). En particular, se ha
demostrado anteriormente un aumento de la presión de transición ası́ como una disminu-
ción de la temperatura de fusión al disminuir el tamaño de partı́cula. En este estudio, se
ha visto que las nanopartı́culas de CdS presentan una transición de fase zinc-blenda (ZB)
→ NaCl en torno a 6 GPa. En el caso de ZnO impurificado con Co2+ se han estudiado
3
los efectos de la transición de fase empleando el Co2+ como sonda local. La transición
de fase irreversible wurtzita (W) → NaCl en nanocristales de Zn1−x Cox O tiene lugar
progresivamente y se completa a 14-15 GPa.
4
Chapter 1
Introduction
Since the famous talk by R. Feynman in 1959, There’s plenty of room at the bottom, a huge
number of research works have been devoted to the fundamental understanding and possi-
ble applications of materials with sizes in the nanometer regime. Spectroscopy deals with
the study of the absorption, reflection, emission or scattering of electromagnetic radiation
by matter. The first spectroscopic study can be attributed to I. Newton’s observation of
the sunlight containing all the colors of the rainbow in 1672. The double-slit experiment
carried out by T. Young in 1801 demonstrated the wave nature of radiation. Moreover,
the concept of photon was developed by A. Einstein to explain some experimental re-
sults which did not fit the wave model. Since the appearance of quantum mechanics,
many spectra of atoms, molecules and solids have been correctly interpreted. Nowadays,
optical spectroscopy represents a powerful and non-destructive tool to understand solid
state physics, this is, to investigate all kind of materials, in particular, nanocrystalline
materials.
The aim of this thesis is two-fold; firstly, this work has been devoted to the study of the
microscopic origin of the optical properties of diverse transition metal (TM) and rare-earth
(RE) ions in different insulating host lattices. Concretely, one of the main goals has been to
establish the mechanisms responsible for the upconversion (UC) luminescence properties
of the studied materials, as well as the structural requirements to increase their quantum
efficiency. Different synthesis methods and impurity concentrations have been compared.
Secondly, phase transitions under high pressure in semiconductor nanocrystals or quantum
dots (QDs) have been studied by means of optical absorption, Raman spectroscopy and
5
6
X-ray diffraction (XRD).
This work has been performed within the High Pressure and Spectroscopy group at
the University of Cantabria. Our research group has a wide experience in the study
of absorption, excitation, luminescence and Raman phenomena as well as in the use of
low temperature and high pressure complementary techniques. Besides, different crystal
growth methods such as the Bridgman technique had been previously implemented in
our laboratories. Nevertheless, this thesis work has been the first one within our research
group involved in the synthesis and optical characterization of nanocrystalline systems.
In this sense, a big effort has been made in order to develop the methods to synthesize
nanoparticles, including different scientific short stays in other research groups, and also in
order to overcome some problems and changes in properties related to the increase of the
surface-to-volume ratio with decreasing size. It is important to point out that synthesizing
the samples ourselves instead of obtaining them from other laboratories allows us to have
a better control of the materials under investigation.
This report is divided in seven chapters. After the Introduction, in Chapter 2 the
relevant theoretical aspects in order to understand the observed optical phenomena are
detailed. The similarities and differences of RE and TM ions and the transitions re-
sponsible of their spectra are outlined. Besides, the quantum confinement effects with
decreasing particle size observed in semiconductor nanocrystals are described. In Chap-
ter 3 the different synthesis methods utilized for the preparation of the nanocrystalline and
microcrystalline samples are exhaustively explained. In Chapter 4 the experimental tech-
niques and setups used for the structural characterization of the samples as well as for the
study of their optical properties are presented. Chapters 5 and 6 show the most relevant
results and discussion of the systems object of this study. Concretely, Chapter 5 is de-
voted to the study of insulating materials doped with RE and TM ions. UC luminescence
phenomena has been observed and characterized in diverse RE3+ -doped nanoparticles
such as Er3+ -Yb3+ co-doped Y2 O3 and NaYF4 , or Tb3+ -Yb3+ and Eu3+ -Yb3+ co-doped
Gd3 Ga5 O12 (GGG) and Y3 Al5 O12 (YAG). The optical properties of LaMgAl11 O19 (LMA)
co-doped with Mn2+ and Yb3+ microcrystalline powders have also been explored. The
effects of temperature and Mn2+ and Yb3+ concentrations on UC luminescence properties
1. Introduction 7
have been studied. Finally, the dependence of GGG: Cr3+ nanoparticles luminescence on
pressure and temperature has been analyzed. In Chapter 6 absorption and Raman spectra
as well as XRD patterns have been studied under high pressure in CdS and Zn1−x Cox O
QDs. CdS nanoparticles show a zinc-blende (ZB) to rock-salt (RS) phase transition at
around 6 GPa, while the phase transition of wurtzite (W) Zn1−x Cox O nanocrystals to the
RS structure takes place gradually and is completed at ca. 14-15 GPa. Finally, the main
conclusions are summarized in Chapter 7.
8
Chapter 2
Theory
2.1 Introduction
The models and approximations utilized for the interpretation of the experimental data are
presented along this chapter. Firstly, the increasing interest of nanomaterials regarding
both fundamental studies and applications is outlined. Secondly, the determination of the
energy levels of an ion in a crystal, and the transitions among these levels, responsible for
the RE and TM ions optical properties, are described. From Fermi’s Golden Rule which
defines the transition probability, concepts like the transition lifetime are introduced.
Non-radiative relaxation processes, responsible for most of the luminescence quenching,
and the different UC mechanisms, as well as the basic concepts of Raman spectroscopy are
also explained. Moreover, the size dependence of the optical properties in semiconductor
nanoparticles is deduced. Finally, structural phase transitions are briefly described.
2.2 Interest of nanoparticles
Remarkable examples of nanotechnology throughout history are the famous Lycurgus Cup
which shows dichroism due to the interaction of light with metal nanoparticles embedded
in a glass matrix, or the addition of metallic constituents to glass, in order to create
stained glass windows [1]. Nowadays, nanoscience and nanotechnology deal with the study
of materials with size in the nanometer range, which is extremely interesting for several
reasons. Firstly, relevant physical laws change at the nanometric scale, this is, many of the
classical laws are not valid anymore and quantum mechanics must be applied. Besides
9
10 2.2. Interest of nanoparticles
quantum confinement effects [2], changes with size reduction are related to the higher
surface-to-volume ratio in nanoparticles [3], [4]. Secondly, some metastable phases are only
observed in reduced-size materials [5]. Thirdly, diverse areas of technology have pointed
their attention in the last years to make smaller components not only for electronics but
also in other fields like new materials, health-sciences or renewable energy sources [1], [6].
The well-known Moore’s law summarizes the scale to get the same function by making the
working elements smaller. The tunnel diode is the first example in which the appearance
of quantum physics at small sizes led to a new device. The fact that the energy gap of QDs,
and therefore the emitting wavelength, is blue-shifted when particle size decreases, is of
great interest for solar cells and bio-labeling [7], [8]. Carbon black which contains carbon
nanotubes, has been used for a long time as an additive to the rubber in automobile tires.
Other examples in which the industry has taken advantage of nanomaterials are titanium
dioxide pigment in white paint, large synthetic biomolecules or nanoparticle slurries for
polishing.
Moving to our research field, there is a considerable interest in the development of

highly-luminescent nanomaterials for biological applications such as bio-labeling, drug
delivery or diagnostics. Metallic and magnetic nanoparticles with the adequate organic
capping, and core-shell semiconductor nanocrystals, have been widely used in biological
experiments [8], [9], [10]. Water-soluble nanocrystals as well as materials with appropri-
ate optical properties are needed. In this sense, the study of the optical properties of
RE-doped nanoparticles for biological applications is increasing because of lanthanides
properties such as narrow absorption and emission bands, high quantum efficiency, and
long lifetimes [11], [12], [13]. Different nanomaterials find also applications in renewable
energy sources such as photovoltaic solar cells technology [7], [14]. UC and downconversion
processes in nanoparticles doped with RE or TM ions may also be implemented for this
purpose. For instance, NaYF4 : Er3+ phosphors have been proposed for the enhancement
of solar cells response [15], [16].
2. Theory 11
2.3 Insulating materials. Optical properties

2.3.1 Doped insulators
Insulators tend to be colorless since they are characterized by a transparency range in the
ultraviolet (UV)- visible - near infrared (NIR) range. The sharp drop in the transmission
for λ > 6000 nm is caused by vibrational absorption, while in the UV spectral region, λ <
200 nm, it is due to absorption in the material bandgap. Insulating hosts can present new
optical properties thanks to the presence of dopant impurity ions, such as RE or TM ions
[17]. The properties of the optically active ions may be understood according to the idea
of a coordination complex, ABn . It is a structure consisting of a central cation A bonded
to the n closest neighbors B, called ligands.
2.3.2 Ions in a static crystalline environment
According to a point-charge model, the valence electrons belong to the dopant ion A,
and the effect of the lattice modifies its energy levels by the influence of the ligand ions
B through the electric field that they produce at the site of A. This electrostatic field is
called the crystal-field, and the Hamiltonian can be written as [18], [19], [20]:
H = HF I + HCF (2.1)
HF I being the free ion A Hamiltonian, and HCF the crystal-field Hamiltonian:
HF I = H0 + H 0 + HSO (2.2)
where H0 includes the kinetic and potential energy in the central-field approximation,
due to the nucleus and the inner closed-shell electrons, H 0 corresponds to the Coulomb
interaction among the outer electrons, and HSO describes the interaction between the
magnetic moment of the electron, s, and its angular moment, l, and it is known as the
spin-orbit coupling.
Depending on the relative importance of the different terms, diverse approaches can
be considered in order to diagonalize the Hamiltonian:
12 2.3. Insulating materials. Optical properties
• Weak crystal field : HCF HSO , H 0 , H0 . In this regime the energy levels of the free
ion A are only slightly perturbed by the crystal field. HCF is taken as a perturbation
over the 2S+1 LJ states (J = L + S). It is appropriate to describe the energy levels of
trivalent RE ions, since for these ions the optically active 4f electrons are screened
by the outer filled 5s2 5p6 sub-shells. These electrons partially shield the crystal field
created by the B ions.
• Intermediate crystal field : HSO HCF < H 0 . In this case the spin-orbit interaction
which is weaker than the crystal field, is initially neglected. The crystal field is
2S+1
considered a perturbation on the L terms.
• Strong crystal field : HSO < H 0 < HCF . In this approach the crystal field term
dominates over the other two. The electron-electron and the spin-orbit interaction
are taken into account as a perturbation. This applies to 3d TM ions and 4f-5d
inter-configurational transitions in RE ions.
2.3.3 The configurational coordinate diagram
So far, we have considered that the optical center A is embedded in a static lattice.
However, in a real crystal, the environment of A is not static but dynamic. In order
to understand the effect of the crystalline vibrations on the optical spectra, we must
consider that the optically active ion A is coupled to the vibrating lattice [18]. The
configurational coordinate model (Fig. 2.1) represents the energy curves in the ground
and excited electronic states as a function of an effective normal mode coordinate. In the
simplest case, this model which can account for such a type coupling, is based on several
approximations; the Franck-Condon approximation, the harmonic approximation and the
fact that only one vibration mode is considered [20].
A = −KQ0 is the electron-lattice coupling constant, where K is the force constant,
considered equal for the ground and excited states for simplicity, and Q0 is the minimum
of the excited state parabola. The sign of A determine whether the equilibrium distance
is longer in the excited state or in the ground state, and it can be obtained from the
pressure dependence of the absorption or emission spectra [21]. The shift between the
2. Theory 13
Figure 2.1: Single configurational coordinate diagram for the AB6 center.
ground and excited states configurational curves is larger for bigger values of A. This
shift which causes a displacement between emission and absorption maxima, is usually
quantified by the Huang-Rhys parameter, S, defined as:
A2 KQ20
S= = (2.3)
2K~ω 2~ω
S is a measure of the linear coupling between the electronic excited state and the
vibrational mode, and it is related with the different equilibrium geometry between ground
and excited states, with S 1 for weak coupling, S between 1 and 4 for medium coupling
and S > 4 for strong coupling. The energy difference between the maximum of the
absorption and emission bands, ESS , is called Stokes shift:
ESS = 2S~ω (2.4)
Then, from the experimental Stokes shift, the Huang-Rhys parameter, and therefore
the distortion of the excited state with respect to the fundamental one, can be obtained.
The stronger the linear coupling is, the larger the bandwith, the Huang-Rhys parameter
and the Stokes shift are. f-f transitions in RE3+ ions show weak coupling with the lattice
due to the mentioned shield of the 4f electrons, hence, the configurational energy curves
are one above the other, and optical spectra consist of fine lines due to pure electronic
transitions with small Stokes shift. On the contrary, d-d transitions in TM ions may also
show strong coupling giving rise to broad bands in their spectra.
2.3.4 Light absorption and emission processes
Determining the energy levels of an optically active center is important because optical
spectra result from transitions among these energy levels. Different processes can take
place such as absorption, due to transitions from lower energy levels to higher energy
levels, and emission, in which the final state is lower in energy than the initial one.
Considering a two-level system consisting of an initial state |i > and a final state |f >
separated by an energy Ef − Ei = ~ωf i , the probability of inducing an optical transition
from the state |i > to the state |f > upon the absorption of a photon is given by the
Fermi’s Golden Rule [20]:
2π
Wi→f = | < i|Hint |f > |2 ρk (ωf i ) (2.5)
~2
where Hint is the radiation-matter interaction Hamiltonian and ρk (ωf i ) is the density
of radiation field modes per unit angular velocity. The parity and spin selection rules
governing a transition are determined by the matrix element < i|Hint |f >.
Relaxation from excited states can take place by both radiative and non-radiative
processes. Purely radiative |f >→ |i > stimulated emission presents the same transition
probability as the inverse absorption process, Wf →i = Wi→f [22]. When spontaneous
emission is the radiative desexcitation channel between |f > and |i > states, the transition
probability is given by the Einstein coefficient, Wfsp→i = Af i . The rate at which the
population of the |f > state, Nf , decays by spontaneous emission in the absence of non-
radiative processes is given by:
dNf
= −Nf Af i (2.6)
dt
t
Nf (t) = Nf (0) exp −
τrad
1
where τrad is the radiative lifetime of the spontaneous emission, τrad = Af i
.
2. Theory 15
Figure 2.2: Classical interpretation of the activation energy, Ea , in non-radiative transitions.
2.3.5 Non-radiative transitions
Contrary to RE ions, in which luminescence is a quite common phenomenon even from

multiple excited states, non-radiative relaxation processes usually dominate the radiative
ones in TM systems. A distinction is usually made between non-radiative processes within
a given complex; multiphonon relaxation, and those involving more than one optically
active center; energy transfer processes [18], [22].
• Multiphonon relaxation: In these processes, the electronic excitation is totally or

partially transformed into vibrational energy within a fempto-picoseconds timescale.
Multiphonon relaxation usually occurs down to the lowest energy excited state,
therefore, if luminescence exists, it takes place from this state. This is known as
Kasha’s rule, and RE compounds are exceptions to this rule. The configurational
coordinate model (Fig. 2.2) provides a qualitative explanation to describe the tem-
perature dependence of multiphonon relaxation from an excited state |f > to some
lower state |i > in TM systems. According to a model proposed by Mott, the non-
radiative desexcitation probability is associated to an activation energy, Ea , and is
given by [23]:

Ea
Wnr f →i = A exp − (2.7)
kT
where Ea is the energy difference between the minimum of the excited state Ef (Q)
parabola and the Ei (Q) - Ef (Q) parabolas crossing point (Fig. 2.2). However, mul-
tiphonon relaxation can also be studied quantically considering tunneling processes
between |f > and |i > states.
In the weak coupling case which is characteristic for RE3+ compounds, the transition
rate between |f > and |i > states through multiphonon relaxation can be expressed
as:
1
Wnr f →i (T = 0K) ∝ exp(−βp) = (2.8)
τnr
This is the so-called energy gap’s law, where p is the reduced energy gap and rep-
resents the number of phonons involved in the relaxation process [24]. For f → f
transitions in RE ions, multiphonon relaxation processes are dominant for energy
separation between levels involving p ≤ 6.
• Energy transfer : All or part of the energy of an initially excited center is transferred
to another center and then either emitted or converted into heat by multiphonon
relaxation. In order to simplify the analysis of this kind of mechanisms, a single
energy transfer step between two ions is considered:
S ∗ + A → S + A∗ (2.9)
where S is the sensitizer (initially in an excited state) and A is the activator (in the
ground state). The interaction between the two ions is responsible for the energy
transfer. Transfer via electric dipole-dipole interaction was first described by Förster
and later expanded by Dexter in order to include higher order electromagnetic and
exchange interactions [25], [26]. The dependence of the energy transfer rate on the
S and A separation, R, is R−6 , R−8 and R−10 for the electric dipole-dipole, dipole-
quadrupole and quadrupole-quadrupole mechanisms, respectively. Energy transfer
by exchange interaction requires very short S-A distances, falling off exponentially
with R.
In lanthanide systems, in which there are often several metastable excited states,
cross-relaxation (CR) and UC processes are very common. In the CR mechanism,
part of the high energy excitation of one ion is transferred to the partner which is
typically in the ground state. One of the UC process, as we will see in Section 2.5,
is exactly the reverse.
The temporal evolution of the luminescence from an excited state shows an exponen-
−t
tial behavior in the absence of energy transfer processes, I(t) = I0 exp( τtotal ), where, in
2. Theory 17
general, the total lifetime is due to contributions of both radiative and non-radiative pro-
1 1 1
cesses, τtotal
= τrad
+ τnr . The rate equations model is extremely useful in order to describe
the dynamics when energy transfer processes are involved; the temporal dependence of
the S and A ions population is studied according to a system of coupled differential equa-
tions. This model has been applied in this work, considering an homogeneous impurity
distribution and an average transfer rate, for the fitting of the temporal evolution of the
UC luminescence in order to understand the different UC mechanisms dynamics.
2.3.6 Dimers
When two ions are close enough, they may interact strongly, lose their individual identities,
and become a dimer. As a consequence, the dimer spectroscopic properties are different
from those of the isolated ions. In both exchange (direct overlap of two ions orbitals) and
superexchange (ions coupled through non-magnetic intermediaries) pathways, the inten-
sity of the interaction strongly depends on the distance between the ions in the dimer,
the geometry, and the electronic configuration of the ions. The so-called Goodenough-
Kanamori rules correlate superexchange interactions with the structural properties. The
sign and magnitude of the resulting kinetic superexchange can be predicted by considera-
tion of the symmetry and the electron occupancy of the interacting orbitals on neighboring
ions, in particular, superexchange coupling is very sensitive to the bond angle [27], [28],
[29].
2.4 Raman Spectroscopy

The Raman effect is a light scattering phenomenon by which part of the incident photons
frequency is used to excite vibrational modes of a molecule or solid. Most of the scattered
photons emerge at the same frequency as the incident ones, ω0 , this is elastic or Rayleigh
scattering. On the contrary, inelastic or Raman scattering is a much less probable event
and takes place at new modified frequencies, ω0 ± ωk , where ωk are the frequencies of the
Raman active vibrational modes [22]. Radiation scattered with a frequency lower than
that of the incident light, ω0 − ωk , or Stokes scattering, is more intense than the one at
higher frequency, ω0 + ωk , or anti-Stokes scattering. This light scattering with change of
18 2.4. Raman Spectroscopy
frequency was first observed independently by Raman and Krishnan in liquids [30], [31]
and Mandelstam in quartz [32] in 1928.
In a typical Raman spectrum, the measured frequencies are expressed as the difference
in wavenumbers (cm−1 ) between the incident and scattered photons, regardless of the
excitation wavelength. Rayleigh scattering is a thousand times stronger than Raman
scattering for strong Raman bands. The ratio of the intensities of the anti-Stokes signal
to the Stokes one is related to the thermal population of the excited vibrational levels
and decreases rapidly when the vibrational mode frequency, ωk , increases. Raman spectra
are strongly dependent on temperature [22]. The Raman scattering cross-section for a
semiconductor is given by:
d2 σ ω4
= vV 4s |εs · χ0 · ε0 |hU U + iω (2.10)
dΩdωs c
with
hv|Hint |cihc|Hep |cihc|Hint |vi

|εs · χ0 · ε0 | ≈ (2.11)
(ωg + ωk − ω0 )(ωg − ω0 )
and
~
hU U + iω = (nk + 1)gk (ω) (2.12)
2N ωk
where ωs is the scattered radiation frequency, V is the total volume of the sample, v is
the excited cross-section, c is the light speed, Hint and Hep are the radiation-electron
and electron-phonon interactions, v and c are the intermediate valence band (VB) and
conduction band (CB) states, ωg is the bandgap frequency, ωk are the frequencies of
the Raman active vibrational modes, and ω0 is the frequency of the incident photons.
Moreover, nk is the phonon distribution as a function of temperature, and gk (ω) is the
phonon lorentzian line-shape factor:
1
nk = (2.13)
exp ~ωk
kT
−1
Γk
2π
gk (ω) = (2.14)
Γk 2

(ωk − ω)2 + 2
2. Theory 19
The intensity of Raman scattering is defined by I = P α2 ω04 , where P is the laser

power, and α is the polarizability of the electrons. The basic selection rules for Raman
scattering arises from the change in polarizability. This means that, in contrast to infrared
(IR) absorption, symmetric vibrations give the most intense Raman scattering. In the
case of solid crystalline samples, when radiation interacts with the material, it induces
vibrations through the whole lattice. Longitudinal (L) modes, along the direction of the
radiation propagation, or transverse (T) modes, in a perpendicular direction, occupy a
band of energies in the material. Lower energy lattice modes are called acoustic modes
and labelled LA and TA while the higher energy type are called optic modes, LO and TO
[33].
2.5 Upconversion
In UC processes, the emission of a high energy photon, usually in the visible, takes place
upon the absorption of two or more lower energy photons, in the IR region. These kind
of processes allow the conversion of the incident radiation in high energy light. UC
phenomena was discovered independently by Auzel [34], [35] and Ovsyankin and Feofilov
[36] in 1966. Opposite to second-harmonic generation, UC processes rely on the presence
of real metastable excited states; one intermediate state which acts as a reservoir of
the incident radiation, and at least one higher energy excited state from which emission
occurs. The most relevant UC mechanisms are summarized in Fig. 2.3; the majority of
them involve some combination of absorption and non-radiative energy transfer steps.
Fig. 2.3(a) shows the ground-state absorption/excited-state absorption (GSA/ESA)
mechanism. In the GSA process the active ion is excited to an intermediate level, and it
is then promoted to an upper emitting state by ESA of a second photon. This process is a
single chromophore process and exhibits an immediate decay of the UC luminescence after
the excitation pulse, because both GSA and ESA steps occur during the laser pulse. The
efficiency strongly depends on the oscillator strength of both GSA and ESA processes, as
well as on the resonance of the involved transitions with the excitation energy.
Fig. 2.3(b) describes the energy transfer upconversion (GSA/ETU) process. In this
case, two ions excited by GSA to the intermediate level interact non-radiatively in order
20 2.5. Upconversion
Figure 2.3: Schematic representation of the most important UC mechanisms: GSA/ESA (a), GSA/ETU
(b), cooperative luminescence (c), and cooperative sensitization (d).
to generate one ion in the upper state and the other in the ground state. Since it is a
non-radiative process, ETU can proceed after the pulse, and this is identified as a rise
in the UC luminescence temporal evolution after a short laser pulse excitation. As we
have seen in Section 2.3.5, the efficiency of the ETU mechanism depends on the distance
between the two involved ions. Hence, ETU processes are very sensitive to active ions
concentration.
Fig. 2.3(c) shows the cooperative luminescence. In this case, the involved ions have
no metastable state to emit visible light. However, the simultaneous emission of two IR
photons in two interacting ions give rise to an emitting visible photon of double energy.
This is a very common process in Yb3+ -doped systems [37], [38], [39].
Fig. 2.3(d) describes the cooperative sensitization mechanism in which two excited
sensitizer ions simultaneously transfer their excitation energy to the activator ion. This
ion has no energy level resonant with the excited state of the sensitizer but it presents
an excited state with twice the excitation energy from which emission takes place after
non-radiative energy transfer. Its temporal behavior is expected to be similar to the
GSA/ETU because it also involves an energy transfer step.
It is well known that the number of photons involved in an UC luminescence process
is indicated by the slope of the luminescence intensity versus the power excitation in a
double-logarithmic representation, I ∝ P n . This is only true, however, in the low pump
power regime. Under high power excitation, the power dependence varies with the UC
mechanism nature [40], [41]. Hence, both GSA/ESA and GSA/ETU shows a quadratic
2. Theory 21
power dependence under low power excitation conditions, and linear dependence at high
power. The efficiency of the different UC processes depends on the number of involved
optically active centers. Cooperative transitions involving three ions are about 3-4 orders
of magnitude weaker than transitions between one or two ions states [42].
2.6 Rare-earth ions

Lanthanides or RE elements are those situated in the sixth period of the periodic table
after the lanthanum, from cerium with the outer configuration 5s2 5p6 4f1 5d1 6s2 , to ytter-
bium with configuration 5s2 5p6 4f14 6s2 [20], [18]. These atoms are usually incorporated in
crystals as divalent or trivalent cations. In RE3+ ions, 5d, 6s and some 4f electrons are re-
moved; hence, trivalent ions deal with transitions between sub-levels of the 4fn electronic
configuration covering an energy range from NIR to deep UV [43]. The optically active
4f electrons are shielded by the 5s and 5p outer electrons. Because of this shielding effect,
the valence electrons are weakly affected by the ligand ions in crystals. This situation
corresponds to the weak crystal field case (see Section 2.3.2). The effect of the crystal field
is to produce a slight shift in the energy of the 2S+1 LJ RE3+ states and to cause additional
splitting. This implies that the main features of RE3+ ions spectra are similar comparing
isolated ions to different hosts. On the contrary, RE2+ ions interact with the crystalline
environment due to the presence of 5d electrons, and show f ↔ d intraconfigurational
transitions, strongly dependent on the host crystal symmetry. These f ↔ d transitions
are also observed in some RE3+ ions.
The interpretation of the RE3+ ions optical spectra is based on the so-called Dieke
2S+1
diagram (Fig. 2.4). Dieke et al. determined the energy of the LJ states for the RE3+
ions in LaCl3 up to 40000 cm−1 [44]. Wegh et al. extended the Dieke diagram from 40000
to about 70000 cm−1 for various lanthanide ions [43]. The width of each state indicates
the magnitude of the crystal-field splitting, it is smaller than 1000 cm−1 in all cases. This
value is in general lower than the spin-orbit splitting. The free ion energy levels are given
by the center of gravity of each multiplet. The same spin-orbit terms in an inverse energy
order for the 4fn and 4f14−n configurations is evidenced in Dieke diagram. Besides, a
comparison of the energy levels of Ce3+ and Yb3+ shows clearly the spin-orbit parameter
22 2.6. Rare-earth ions
Figure 2.4: Dieke diagram of RE3+ ions in LaCl3 [44]. Semicircles indicate the emitting levels.
2. Theory 23
increase for larger atomic number.

Since electron-phonon coupling effects are weak, optical transitions between 4fn states
in RE3+ are characterized by sharp lines. These transitions occur between states of the
same parity and then they are expected to occur by a magnetic dipole process. However, a
mixing of 4fn states with opposite-parity states of the 4fn−1 -5d configuration can take place
by coupling with odd vibrational modes, allowing a forced electric dipole component to
the transition. Spin is not a good quantum number to consider in f-f transitions because
of the spin-orbit coupling. Some general selection rules for electric dipole transitions
between RE3+ 4fn states obtained from the Judd-Ofelt theory are |∆J| ≤ 6 with J =
0 ↔ J 0 = 0 forbidden, and J = 0 ↔ odd J 0 much weaker than J = 0 ↔ even J 0 [20].
Interconfigurational 4fn ↔ 4fn−1 5d transitions are parity allowed and usually appear in
the UV region for RE3+ ions. Their energy strongly depends on the local environment
because 5d levels are not so efficiently screened.
2.7 Transition metal ions

TM ions (iron group) are formed from the elements situated in the fourth period of
the periodic table in which the 3d shell is being filled. In ionic solids, such elements
lose the outer 4s electrons and some 3d electrons, and their electronic configurations are
1s2 2s2 2p6 3s2 3p6 3dn , where n < 10 [20]. Since 3d electrons, responsible for the optical
transitions, interact strongly with electric fields of nearby ions, the crystal-field effects
are, in general, more important than for RE3+ ions. Besides, TM ions show strong
electron-lattice coupling; as a result, the spectra can present both broad (S > 0) and
sharp (S ∼ 0.1) bands [18]. According to the strong crystal field scheme, and taking
into account that the electron-electron and crystal-field interactions dominate over the
spin-orbit term, the electronic levels splitting of the TM ion can be obtained according
to each complex symmetry [20]. In an octahedral crystal field, the five d levels are split
into a triply-degenerate level t2g and a doubly-degenerate level eg ; the strength of the
crystal field responsible for this splitting, ∆, is called 10Dq. The energy of the transitions
observed experimentally in the TM spectra can be described by using ∆=10Dq and the
so-called Racah parameters B and C [45].
24 2.7. Transition metal ions
Figure 2.5: Tanabe-Sugano diagram for a d3 ion in an octahedral crystal field or a d7 ion in a tetrahedral
crystal field (left), and Tanabe-Sugano diagram for a d5 ion (right). The vertical lines indicate the crystal
field strength for Co2+ , Cr3+ and Mn2+ ions, in the lattices studied in this work; ZnO, Gd3 Ga5 O12 and
LaMgAl11 O19 , respectively.
Sugano and Tanabe calculated the energy of the states deriving from dn ions (2<
n <8) as a function of the octahedral crystal field. These calculations are represented in
the Tanabe-Sugano diagrams which show the energy of the states in units of B (E/B )
compared to the ground state energy, as a function of the ratio between the crystal-field
strength and the interelectronic interaction, measured in units of ∆/B, for a given C/B
ratio [46]. The energy level scheme of a 3d10−n ion in a tetrahedral crystal field is the same
as in an octahedral crystal field but with negative 10Dq, hence, it has the same pattern
as the splitting of a 3dn system in an octahedral crystal field. The relationship between
the crystal-field strenghts is given by 10Dq(octahedral)= − 49 10Dq(tetrahedral). Figure
2.5 shows the Tanabe-Sugano diagrams for d3 and d5 systems in an octahedral crystal
field. These diagrams are appropriate to describe the TM ions that have been studied
in this work: Cr3+ (3d3 system) in an octahedral crystal field and Co2+ (3d7 system) in
a tetrahedral crystal field (Fig. 2.5(a)), and Mn2+ (3d5 system) in a tetrahedral crystal
2. Theory 25
field (Fig. 2.5(b)). Tanabe-Sugano diagrams allow to predict if an optical band of a TM

ion is narrow or broad from the slope of the involved states. In general, the magnitude of
the slope is proportional to the electron-lattice coupling constant. Therefore, the states
with energies independent of the crystal field give rise to narrow bands while broad bands
are associated with transitions involving large slope energy levels. Despite the fact that
non-radiative processes are more important in TM than in RE systems, UC luminescence
processes have also been observed in different TM ions such as Pb+ , Ti2+ , Mn2+ , Ni2+ or
Cr3+ [47], [48], [49], [50].
2.8 Semiconductor nanostructures
The bandgap of a semiconductor is defined as the energy difference between the highest
energy VB states and the lowest energy CB states. If an electron is excited from the
VB to the CB leaving a hole in the VB, the electron-hole (e-h) pair may form a bound
state through Coulomb interactions. This bound state (the so-called Wannier-Mott or
free exciton) has an energy slightly lower than the bandgap energy. The Bohr radius of
the exciton, aB , is given by the following equation:
4πε0 ε∞ ~2 1 1
aB = = a 0 ε ∞ (2.15)
m0 e2 µ∗ µ∗
where ε0 is the vacuum dielectric constant, ε∞ is the high frequency relative dielectric
constant of the medium, m0 is the mass of the free electron, µ∗ , with 1
µ∗
= 1
m∗e
+ 1
m∗h
,
is the effective reduced e-h mass (m∗e and m∗h are the effective electron and hole masses),
and a0 is the Bohr radius of a hydrogen atom (0.529 Å). Since the effective masses are
significantly smaller than m0 , and ε∞ is much larger than 1, the resulting values for aB
are much larger than a0 , in the range 1-10 nm for the common semiconductors. The Bohr
radius of the exciton is 2.6 nm for bulk CdS in the ZB structure [51] and 0.9 nm for bulk
W-ZnO [52]. Semiconductor nanoparticles exhibit quantum size or quantum confinement
effects when the dimensions of the system approaches the Bohr radius of the exciton, as a
result of spatial confinement of the electrons and holes in the nanoparticles; in these cases,
they are also called QDs [53], [54]. Two manifestations of these quantum size effects are
26 2.9. Interband absorption in semiconductors
the increase in the bandgap with decreasing particle size, and the transition from energy
bands to discrete energy levels. The chemical bonds contraction at the surface due to
coordination deficiency and the rise of the surface-to-volume ratio have been shown to be
responsible for these changes at the nanoscale [55].
Different approaches have been proposed to explain the bandgap expansion of semi-
conductor nanocrystals [55]. The widely accepted model was initially proposed by Efros
[56] and further extended by Brus [53], [57] to include the term of Coulomb interaction of
an e-h pair. The confinement effect on the bandgap, Eg , is then expressed as a function
of the semiconductor nanoparticle radius, R:
π 2 ~2 1 1.8e2
Eg (R) = Eg + − (2.16)
2m0 R2 µ∗ 4πε0 ε∞ R
where the first term, Eg , represents the bandgap of the bulk semiconductor, the second
term represents the spatial confinement of electrons in the CB and holes in the VB by
the potential barrier of the surface or a mono-potential well of the quantum box and
has a 1/R2 dependence, and the third term represents the Coulomb energy with a 1/R
dependence. Generally, the second term is more important than the third one, leading to
an increase in the QDs energy bandgap when reducing the particle size.
2.9 Interband absorption in semiconductors

2.9.1 Introduction
Semiconductors absorption spectra in the visible region are due to interband transitions,
in which an electron is excited from the VB to the CB. A continuous absorption spectrum
starting from the low energy threshold at the energy bandgap, Eg , is observed for bulk
semiconductors. A step-like behavior is expected for quantum wells while a series of
delta functions is predicted for QDs [17]. Although the joint density of states changes
when the QDs size is reduced, the only size effect that has been considered in this work
is the blue-shift of the absorption edge with decreasing particle size. Therefore, the
equations describing interband transitions in bulk semiconductors have been used, and
are shown in the next two sections. The interband absorption rate depends on the band
2. Theory 27
structure of the solid. Considering the relative positions of the CB minimum and the VB
maximum in the Brillouin zone, the bandgap may be direct or indirect. Conservation
of momentum implies that absorption processes are represented by vertical lines on E-k
diagrams. Hence, if the bandgap is indirect, the transition must involve a phonon to
satisfy momentum conservation [17].
Figure 2.6: Interband transitions in semiconductors. Direct bandgap (a) and indirect bandgap (b).
2.9.2 Band edge absorption in direct gap semiconductors
In a direct gap semiconductor, both the VB maximum and the CB minimum occur at the
same value of k (see Fig. 2.6(a)). Assuming a simplified parabolic form for the VB and
CB, the conservation of energy during a transition requires that:
~2 k 2
~ω = Eg + (2.17)
2µ
where µ is the reduced e-h mass. The joint e-h density of states, g(~ω), is then given by:
g(~ω) = 0 For ~ω < Eg (2.18)

3/2
1 2µ
g(~ω) = 2
(~ω − Eg )1/2 For ~ω ≥ Eg
2π ~2
Fermi’s Golden Rule tells us that the absorption rate for a dipole-allowed interband
transition is proportional to the joint density of states. Hence, the following behavior is
expected for the frequency dependence of the absorption coefficient, α(~ω) [17]:
α(~ω) = 0 For ~ω < Eg (2.19)
α(~ω) ∝ (~ω − Eg )1/2 For ~ω ≥ Eg
According to eq. (2.19), a straight line behavior is obtained plotting α2 against the
photon energy, ~ω, in the spectral region close to the bandgap. The bandgap is the point
at which the absorption goes to zero.
2.9.3 Band edge absorption in indirect gap semiconductors
Indirect gap materials have their CB minimum away from the Brillouin zone center, as
shown schematically in fig. 2.6(b). A photon can excite an electron from the VB to the CB
involving a phonon to conserve momentum, k. Conservation of energy and momentum
requires that:
~ω = ECV ± ~Ω (2.20)
kV − kC = ±Q (2.21)
where ECV is the difference in energy between VB and CB, ~Ω is the phonon energy, kV
and kC are the respective wave vectors of the involved states in VB and CB, and Q is the
phonon wave vector. The ± sign represents the phonon absorption or emission. Indirect
transitions are second-order processes, the transition rate is therefore much smaller than
for direct absorption. The frequency dependence of the indirect absorption coefficient in
this case is given by:
αi (~ω) ∝ (~ω − Eg ∓ ~Ω)2 (2.22)
If α1/2 is plotted against ~ω, the indirect gap semiconductor data fits well to a straight
line. The experimental data also shows a tail caused by higher frequency phonons or
multiphonon absorption.
It must be pointed out that since the frequency dependence is different for direct and
indirect gap semiconductors, it provides a very convenient way to determine the nature
of the bandgap.
2. Theory 29
2.10 Structural phase transitions. Equations of state
The crystalline structure of a solid may undergo modifications when pressure or temper-
ature conditions are changed. The different phases of a given material together with its
P − T stability points compose the phase diagram of the material. From a thermody-
namic point of view, a given phase is stable when the Gibbs free energy is minimum.
First-order phase transitions exhibit a discontinuity in the first derivative of the Gibbs
free energy. A diminution of the volume or the entropy is experimentally observed at
the transition point. Transitions showing hysteresis in both pressure and temperature are
first-order transitions. Second-order phase transitions are continuous in the first deriva-
tive but exhibit discontinuity in a second derivative of the free energy, they take place at
well-defined temperature and pressure, and they present no hysteresis [58]. It is impor-
tant to point out that pressure-induced phase transitions lead to transformations to more
compact structures, increasing the coordination number and involving volume reduction.
Density increases and, therefore, they are usually first-order phase transitions.
The appearance of a structural phase transition may have important consequences

not only in the structural characterization measurements, but also in the optical proper-
ties. The absorption or emission spectra of an optically active ion change drastically if
the new phase involves modifications of the coordination number or bond distances. Be-
sides, changes in the bandgap absorption and Raman spectra of semiconductor materials,
associated to the transformation into a new crystalline structure are also very common.
An equation of state provides a relationship between two or more state variables, such
as temperature, pressure or volume. For a given phase, temperature, volume, and pressure
are not independent quantities; they are connected by a relationship of the general form
f (P, V, T ) = 0. For first-order phase transitions, each structural phase in the P − T
diagram has a characteristic equation of state.
Considering an isothermal relation between P and V , the simplest reasonable model

∂P

for an equation of state would be with a constant bulk modulus, B0 = −V ∂V T
. A more
sophisticated equation of state was derived by Murnaghan assuming that isothermal B is a
linear function of P , B = B0 + B00 · P . The bulk modulus pressure derivative, B 0 = ∂B

∂P T
,
30 2.10. Structural phase transitions. Equations of state
is considered constant, B 0 = B00 , and therefore [59]:
−1/B00
B0

V (P ) = V0 1 + 0P (2.23)
B0
V0 being the equilibrium volume, V (P = 0). This equation is appropriate for compressions
lower than 15%, however, for higher compressions more elaborated equations of state are
needed, such as Birch-Murnaghan equation [60] or Rose-Vinet equation [61].
Bibliography
[1] G.L. Hornyak, J. Dutta, H.F. Tibbals, and A.K. Rao, editors. Introduction to
Nanoscience. CRC Press, Taylor and Francis group, Florida, 2008.
[2] A. Trave, F. Buda, and A. Fasolino. Band-Gap Engineering by III-V Infill in Sodalite.
Phys. Rev. Lett., 77: 5405–5408, 1996.
[3] J.R. Agger, M.W. Anderson, M.E. Pemble, O. Terasaki, and Y. Nozue. Growth of
Quantum-Confined Indium Phosphide inside MCM-41. J. Phys. Chem. B, 102: 3345–
3353, 1998.
[4] F. Vetrone, J.C. Boyer, J.A. Capobianco, A. Speghini, and M. Bettinelli. A spec-
troscopic investigation of trivalent lanthanide doped Y2 O3 nanocrystals. Nanotech.,
15: 75–81, 2004.
[5] C. Ricolleau, L. Audinet, M. Gandais, and T. Gacoin. Structural transformations in

II-VI semiconductor nanocrystals. Eur. Phys. J. D, 9: 565–570, 1999.
[6] E.L. Wolf, editor. Nanophysics and Nanotechnology. Wiley-VCH, Weinheim, Ger-
many, 2004.
[7] M. Stupca, M. Alsalhi, T. Al Saud, A. Almuhanna, and M.H. Nayfeh. Enhancement

of polycrystalline silicon solar cells using ultrathin films of silicon nanoparticle. Appl.
Phys. Lett., 91: 063107, 2007.
[8] M. Bruchez Jr, M. Moronne, P. Gin, S. Weiss, and A.P. Alivisatos. Semiconductor
Nanocrystals as Fluorescent Biological Labels. Science, 281: 2013–2016, 1998.
31
32 BIBLIOGRAPHY
[9] R. Elghanian, J.J. Storhoff, R.C. Mucic, R.L. Letsinger, and C.A. Mirkin. Selective
Colorimetric Detection of Polynucleotides Based on the Distance-Dependent Optical
Properties of Gold Nanoparticles. Science, 277: 1078–1081, 1997.
[10] W.C.W. Chan and S. Nie. Quantum Dot Bioconjugates for Ultrasensitive Nonisotopic
Detection. Science, 281: 2016–2018, 1998.
[11] M. Nyk, R. Kumar, T.Y. Ohulchanskyy, E.J. Bergey, and P.N. Prasad. High Contrast
in Vitro and in Vivo Photoluminescence Bioimaging Using Near Infrared to Near
Infrared Up-Conversion in Tm3+ and Yb3+ Doped Fluoride Nanophosphors. Nano
Lett., 8: 3834–3838, 2008.
[12] F. Vetrone and J.A. Capobianco. Lanthanide-doped fluoride nanoparticles: lumi-

nescence, upconversion, and biological applications. Int. J. Nanotech., 5: 1306–1339,
2008.
[13] C. Vancaeyzeele, O. Ornatsky, V. Baranov, L. Shen, A. Abdelrahman, and M.A.

Winnik. Lanthanide-Containing Polymer Nanoparticles for Biological Tagging Appli-
cations: Nonspecific Endocytosis and Cell Adhesion. J. Am. Chem. Soc., 129: 13653–
13660, 2007.
[14] P.V. Kamat. Meeting the Clean Energy Demand: Nanostructure Architectures for
Solar Energy Conversion. J. Phys. Chem. C, 111: 2834–2860, 2007.
[15] A. Shalav, B.S. Richards, and T. Trupke. Application of NaYF4 : Er3+ up-converting
phosphors for enhanced near-infrared silicon solar cell response. Appl. Phys. Lett.,
86: 013505, 2005.
[16] L. Aarts, B.M. van der Ende, and A. Meijerink. Downconversion for solar cells in
NaYF4 : Er, Yb. J. Appl. Phys., 106: 023522, 2009.
[17] M. Fox, editor. Optical Properties of Solids. Oxford University Press, Oxford, 2001.
[18] J. Garcı́a Solé, L.E. Bausá, and D. Jaque, editors. An Introduction to the Optical
Spectroscopy of Inorganic Solids. John Wiley and Sons Ltd, England, 2005.
BIBLIOGRAPHY 33
[19] J.S. Griffith, editor. The Theory of Transition-Metal Ions. Cambridge University
Press, London, 1971.
[20] B. Henderson and G.F. Imbusch, editors. Optical Spectroscopy of Inorganic Solids.
Clarendon Press, Oxford, 1989.
[21] R. Valiente, F. Rodrı́guez, J. González, H.U. Güdel, R. Martı́n-Rodrı́guez, L. Nataf,

M.N. Sanz-Ortiz, and K. Krämer. High pressure optical spectroscopy of Ce3+ -doped
Cs2 NaLuCl6 . Chem. Phys. Lett., 481: 149–151, 2009.
[22] E.I. Solomon and A.B.P. Lever, editors. Inorganic Electronic Structure and Spec-
troscopy, volume I. John Wiley and Sons, New York, 1999.
[23] N.F. Mott. Proc. Roy. Soc. (London), A 167: 384, 1938.
[24] J.M.F. van Dijk and M.F.H. Schuurmans. On the nonradiative and radiative decay
rates and a modified exponential energy gap law for 4f−4f transitions in rare-earth
ions. J. Chem. Phys., 78: 5317–5323, 1983.
[25] T. Förster. Ann. der Physik, 2: 55, 1948.
[26] D.L. Dexter. J. Chem. Phys., 21: 836, 1953.
[27] J.B. Goodenough, editor. Magnetism and the Chemical Bond. Interscience-Wiley,
New York, 1963.
[28] J. Kanamori. J. Phys. Chem. Solids, 10: 87, 1959.
[29] H. Weihe and H.U. Güdel. Quantitative Interpretation of the Goodenough-Kanamori

Rules: A Critical Analysis. Inorg. Chem., 36: 3632–3639, 1997.
[30] C.V. Raman and K.S. Krishnan. A New Type of Secondary Radiation. Nature,
121: 501–502, 1928.
[31] C.V. Raman. A Change of Wave-length in Light Scattering. Nature, 121: 619–619,
1928.
34 BIBLIOGRAPHY
[32] G.S. Landsberg and L.J. Mandeltsam. Eine neue Erscheinung bei der Lichtzerstreu-
ung in Krystallen. Naturwissenschaften, 16: 557–558, 1928.
[33] W.E. Smith and G. Dent, editors. Modern Raman Spectroscopy - A Practical Ap-
proach. John Wiley and Sons, England, 2005.
[34] F. Auzel. Acad. Sci., 262: 1016, 1966.
[35] F. Auzel. Acad. Sci., 263 B: 765, 1966.
[36] V.V. Ovsyankin and P.P. Feofilov. Jetp. Lett., 3: 322, 1966.
[37] E. Nakazawa and S. Shionoya. Cooperative Luminescence in YbPO4 . Phys. Rev.

Lett., 25: 1710–1712, 1970.
[38] P. Goldner, F. Pellé, D. Meichenin, and F. Auzel. Cooperative luminescence in

ytterbium-doped CsCdBr3 . J. Lumin., 71: 137–150, 1997.
[39] M.P. Hehlen and H.U. Güdel. Optical spectroscopy of the dimer system Cs3 Yb2 Br9 .
J. Chem. Phys., 98: 1768–1775, 1993.
[40] M. Pollnau, D.R. Gamelin, S.R. Lüthi, H.U. Güdel, and M.P. Hehlen. Power depen-
dence of upconversion luminescence in lanthanide and transition-metal-ion systems.
Phys. Rev. B, 61: 3337–3346, 2000.
[41] J.F. Suyver, A. Aebischer, S. Garcı́a-Revilla, P. Gerner, and H.U. Güdel. Anomalous
power dependence of sensitized upconversion luminescence. Phys. Rev. B, 71: 125123,
2005.
[42] F. Auzel. Upconversion and Anti-Stokes Processes with f and d ions in Solids. Chem.
Rev., 104: 139–173, 2004.
[43] R.T. Wegh, A. Meijerink, R.J. Lamminmäki, and J. Hölsä. Extending Dieke’s dia-
gram. J. Lumin., 87-89: 1002–1004, 2000.
[44] G.H. Dieke, editor. Spectra and Energy Levels of Rare Earth Ions. Wiley Interscience,
New York, 1968.
BIBLIOGRAPHY 35
[45] S. Sugano and Y. Tanabe. J. Phys. Soc. Jpn., 9: 753, 1954.
[46] S. Sugano, Y. Tanabe, and H. Kamimura, editors. Multiplets of transition metal ions
in solids. Academic Press, New York, 1970.
[47] J.M. Spaeth, R.H. Bartram, M. Rac, and M. Fockele. Upconversion by excited state
absorption of Pb+ (l) centres in alkaline-earth fluorides. J. Phys.: Condens. Matter,
3: 5013–5022, 1991.
[48] R. Valiente, O.S. Wenger, and H.U. Güdel. New photon upconversion processes in
Yb3+ doped CsMnCl3 and RbMnCl3 . Chem. Phys. Lett., 320: 639–644, 2000.
[49] S. Garcı́a-Revilla, P. Gerner, H.U. Güdel, and R. Valiente. Yb3+ -sensitized visi-
ble Ni2+ photon upconversion in codoped CsCdBr3 and CsMgBr3 . Phys. Rev. B,
72: 125111, 2005.
[50] S. Heer, M. Wermuth, K. Krämer, and H.U. Güdel. Sharp 2 E upconversion lumines-
cence of Cr3+ in Y3 Ga5 O12 codoped with Cr3+ and Yb3+ . Phys. Rev. B, 65: 125112,
2002.
[51] Y. Kayanuma. Quantum-size effects of interacting electrons and holes in semicon-

ductor microcrystals with spherical shape. Phys. Rev. B, 38: 9797–9805, 1988.
[52] V.A. Fonoberov and A.A. Balandin. Radiative lifetime of excitons in ZnO nanocrys-
tals: The dead-layer effect. Phys. Rev. B, 70: 195410, 2004.
[53] L.E. Brus. Electron-electron and electron-hole interactions in small semiconductor

crystallites: The size dependence of the lowest excited electronic state. J. Chem.
Phys., 80: 4403–4409, 1984.
[54] P.F. Trwoga, A.J. Kenyon, and C.W. Pitt. Modeling the contribution of quantum
confinement to luminescence from silicon nanoclusters. J. Appl. Phys., 83: 3789–3794,
1998.
[55] C.Q. Sun, T.P. Chen, B.K. Tay, S. Li, H. Huang, Y..B Zhang, L.K Pan, S.P. Lau,
and X.W. Sun. An extended quantum confinement theory: surface-coordination
36 BIBLIOGRAPHY
imperfection modifies the entire band structure of a nanosolid. J. Phys. D: Appl.

Phys., 34: 3470–3479, 2001.
[56] Al.L. Efros and A.L. Efros. Interband absorption of light in a semiconductor sphere.
Sov. Phys. Semicond., 16: 772–775, 1982.
[57] L.E. Brus. On the development of bulk optical properties in small semiconductor
crystallites. J. Lumin., 31-32: 381–384, 1984.
[58] R. Ehrenfest. Proc. Acad. Sci. Amsterdam, 36: 153–157, 1933.
[59] F.D. Murnaghan. The Compressibility of Media under Extreme Pressures. P. Natl.
Acad. Sci., 30: 244–247, 1944.
[60] F. Birch. Finite Elastic Strain of Cubic Crystals. Phys. Rev., 71: 809–824, 1947.
[61] P. Vinet, J. Ferrante, J.R. Smith, and J.H. Rose. A universal equation of state for
solids. J. Phys. C: Solid State Phys., 19: L467–L473, 1986.
Chapter 3
Synthesis
3.1 Introduction
In this chapter, the different synthesis methods used for the preparation of the samples
are described. Er3+ and Yb3+ co-doped Y2 O3 and NaYF4 nanoparticles, as well as CdS
nanopowders were prepared in a planetary ball mill at the Solid State Physics laboratory of
the University of Cantabria. RE-doped Y2 O3 nanocrystals were also synthesized following
a combustion reaction at the Inorganic Chemistry laboratory of our university. Tb3+ -
Yb3+ and Eu3+ -Yb3+ co-doped ternary oxides GGG and YAG nanocrystalline samples,
and Cr3+ -doped GGG nanoparticles were grown during a two month stay within Prof.
Marco Bettinelli’s group at the Biotecnology Department, University of Verona. All of
them were obtained using Pechini’s method [1]. Some of these samples were coated with
SiO2 using Stöber method [2]. As another result of this collaboration, microcrystalline
powders of LMA doped with Mn2+ and Yb3+ were also synthesized in one of Prof. Marco
Bettinelli’s laboratory. Colloidal Zn1−x Cox O nanocrystals were prepared at the Chemistry
Department, University of Washington, thanks to a collaboration with Prof. Daniel R.
Gamelin during a three month stay in his group.
3.2 Mechano-chemical processes in a planetary ball

mill
A wide variety of processes such as crystallization of amorphous alloys, particle size re-
duction or mechano-chemical reactions can take place in a planetary ball mill [3]. In all
37
38 3.2. Mechano-chemical processes in a planetary ball mill
cases, mechanical ball milling consists of repeated fracture, mixing and welding of pow-
der particles (less than 10 µm in size). When the mill rotates in one direction the vials
rotate in the opposite one, and there are collisions among the powders, the balls and the
vials walls [4]. The occurrence of these mechano-chemical reactions is attributed to the
heat generated in the milling process, favored by the large area of contact between the
solids [5]. At the point of collision, the solids deform and even melt, forming hot points
where the molecules can reach very high vibrational excitation leading to bond breaking.
Related to the final nanoparticles size, it is important to point out the existence of a
steady-state grain size during high-energy ball-milling. If the initial size is bigger than
the steady-state size, the crystallite size decreases during the milling, but smaller particles
show an increase in size reaching finally such steady-state. This can be explained by an
equilibrium between fracturing and crystal growth processes during the milling [6].
Figure 3.1: Planetary ball mill used for the synthesis of Y2 O3 , NaYF4 and CdS nanoparticles at the
Solid State Physics laboratory, University of Cantabria.
The high-energy ball-milling synthesis was carried out using a planetary ball mill
(Retsch PM 400/2) (Fig. 3.1). The most important parameters to take into account in
these processes are: Vials and balls material, balls size, atmosphere under which synthesis
is performed (Air or argon), total milling time, milling and pause times, angular velocity
(rpm), ball to sample mass ratio and the addition of a diluent. Changes in every parameter
are decisive not only for the obtention of the desired phase but also for the final size of the
nanocrystals. Searching for the optimum conditions for each synthesis process involved
an arduous and systematic work. In some cases, the ball mill was also used in order to
3. Synthesis 39
homogenize the size of nanoparticles prepared using other synthesis methods.
3.2.1 Er3+ , Yb3+ co-doped Y2 O3
Y2 O3 : Er3+ , Yb3+ nanoparticles were obtained in a planetary ball mill by intimately

mixing stoichiometric quantities of yttrium, erbium and ytterbium oxides according to
the reaction:
(1 − A − B)Y2 O3 + AEr2 O3 + BYb2 O3 → Y2 O3 : A%Er3+ B%Yb3+ (3.1)
The following parameters were used:
• ZrO2 vials and balls, with ball diameter of 10 mm. Other materials such as steal
had been previously tried, but they contaminated the samples.
• Ball to powder mass ratio of 10:1, with a total amount of 6 g.
• Air atmosphere and 200 rpm.
• 5 minute stop every 30 minutes of milling. The dependence of the particle size on
the total milling time was studied in order to determine the most advantageous
conditions.
3.2.2 Er3+ , Yb3+ co-doped NaYF4
Different procedures have been previously described in the literature for the synthesis of
sodium yttrium fluoride, NaYF4 , such as heating the mixture NaF + YF3 at 800 ◦ C under
HF gas [7] or a hydrothermal method that requires complicated apparatus [8]. These re-
strictions in fluoride synthesis have limited their development in comparison, for instance,
with oxides. Fluorides have also been prepared by mechano-chemical synthesis [9]. This
fact, and the difficulty in synthesizing the pure hexagonal phase by other methods, led
us to prepare hexagonal NaYF4 : Er3+ , Yb3+ nanocrystals using the planetary ball mill
starting from stoichiometric quantities of fluorides according to the reaction:
NaF + (1 − A − B)YF3 + AErF3 + BYbF3 → NaYF4 : A%Er3+ , B%Yb3+ (3.2)

Bearing in mind the necessity of using parameters which increase collisions energy
during the milling [9], it is convenient to increase the ball to powder mass ratio or the
angular velocity. A systematic study to obtain the optimal synthesis conditions was
accomplished; different grinding times, velocity and mass ratio were tried. The final
parameters established as the most convenient for the synthesis of hexagonal NaYF4
were:
• ZrO2 vials and 20 balls.
• Ball to sample mass ratio of 30:1.
• Argon atmosphere and 300 rpm.
• 2 hours stop every grinding hour.
Different total milling times were used in order to get the pure hexagonal phase.
Changes in any of the former parameters are rather critical for the obtention of the desired
hexagonal phase. For example, if a ball to powder mass ratio of 20:1 is used, cubic phase
is obtained after 5 hours milling. If a 5 minute pause is set, peaks corresponding to cubic
phase are more intense than hexagonal peaks after 8 hours milling.
3.2.3 Pure and Yb3+ -doped CdS
Pure and Yb3+ -doped CdS nanoparticles were prepared according to the following mechano-
chemical reaction in a planetary ball mill [10], [11]:
CdCl2 + YbCl3 + Na2 S + 10NaCl → CdS:Yb3+ + 12NaCl (3.3)
The starting materials were CdCl2 , YbCl3 , Na2 S and NaCl. Milling was performed
under high purity argon atmosphere using hardened steel vials and balls, a ball to sample
mass ratio of 20:1 and 300 rpm. NaCl was used as diluents in order to prevent nanopar-
ticles aggregation and has little effect on the obtained particles size. The particle size
of the product depends on the size of the starting reactants; smaller Na2 S particles give
rise to smaller CdS nanocrystals [12]. In this sense, the mixture Na2 S + 10NaCl was ini-
tially milled for 1 hour. Then, CdCl2 was added and the mixture was milled for another
3. Synthesis 41
hour. The as-obtained powders were washed with distilled water, filtered and dried under
vacuum in order to remove NaCl contents.
3.3 Combustion reaction

Y2 O3 nanocrystals doped with different Er3+ and Yb3+ concentrations were synthesized
following the glycine-nitrate solution combustion reaction:
3 5 25
3M(NO3 )3 + 5NH2 CH2 COOH + 9O2 → M2 O3 + N2 + 9NO2 + 10CO2 + H2 O (3.4)
2 2 2
where M = Y3+ , Er3+ or Yb3+ [13], [14], [15]. The experimental procedure was as follows:
yttrium, ytterbium, erbium nitrates and glycine were mixed in an appropriate ratio and
dissolved in distilled water to form the precursor solution. This solution was concentrated
by heating in a quartz crucible or in a pyrex beaker on a heating plate. When excess
water was evaporated, a spontaneous ignition took place and the combustion reaction
was finished originating the oxide nanoparticles (fig. 3.2).
Figure 3.2: Different sequences during the combustion synthesis of RE-doped Y2 O3 nanoparticles at
the Inorganic Chemistry laboratory, University of Cantabria.
The combustion temperature has a great influence on the final particle size, and this
reaction temperature can be controlled by changing the glycine-to-nitrates ratio (G/N).
For the preparation of RE oxides, it has been reported that lower temperature leads
to smaller size for the obtained nanoparticles [15]. In this sense, a G/N lower than
stoichiometric, G/N=1, was used in order to get as small nanoparticles as possible.
In the described process, Y2 O3 nanocrystals adsorb part of the H2 O and CO2 molecules
produced during the reaction. The presence of these groups on the nanoparticles surface
42 3.4. Pechini’s method (sol-gel)
yields higher energy phonons which make multiphonon relaxation much more probable,
and are responsible for the reduction in the luminescence (see Section 5.2.1). Different
heat treatments were carried out in order to reduce the amount of OH− and CO2−
3 ions
which were analyzed by IR absorption spectroscopy. All samples were fired at different
temperatures for different periods of time. It must be considered that longer treatments at
higher temperatures reduce surface contamination, but they can also induce nanoparticles
aggregation with the resulting increase in size.
3.4 Pechini’s method (sol-gel)

GGG and YAG nanocrystalline powders co-doped with Tb3+ or Eu3+ and Yb3+ , and
Cr3+ -doped GGG nanoparticles have been prepared using the sol-gel Pechini’s method as
described in Ref. [16]. Among the sol-gel procedures, the preparation of oxides using the
formation of an intermediate precursor polyester was proposed by Pechini in 1960’s [1].
This technique was initially proposed to prepare niobates and titanates ceramic materials
or thin films. Pechini’s method is based on the ability of some α-hydroxycarboxylic acids,
such as citric acid (CA), to form chelates between the metal ions (Gd, Ga, Y or Al in
our case) and the acid. When this complex is mixed with some polyhydroxyl alcohol,
usually polyethylene glycol (PEG), and is heated, the polymerization process takes place.
Finally, the polymeric matrix is thermally decomposed to remove the organic groups and
form the stoichiometric oxide phase [17].
In our case, the synthesis of ternary oxides according to this method was as follows:
First, the precursor solution was prepared by dissolving the nitrates in distilled water
and adding a suitable amount of CA, according to the ratio CA moles = 2 × Nitrates
moles. The starting materials were reagent grade Gd(NO3 )3 , Ga(NO3 )3 and Ln(NO3 )3 or
Cr(NO3 )3 for the GGG samples and Y(NO3 )3 , Al(NO3 )3 and Ln(NO3 )3 in the case of the
YAG samples. Then, the solution was heated at 90 ◦ C under stirring until a transparent
solution was obtained. At this step, molecules between metal ions and acid are formed.
Afterwards, PEG was added to the solution according to PEG moles = CA moles / 20
and it was stirred for 15 minutes so that polymerization could take place. Finally, the
obtained sol was heated at 90 ◦ C for 24 hours in order to form the gel, an amorphous
3. Synthesis 43
mixture of oxides, and this gel was fired at 800 ◦ C for 16 hours to induce its crystallization
and to form GGG and YAG nanoparticles (Fig. 3.3).
Figure 3.3: Sol, gel, and powder steps during the GGG and YAG nanoparticles synthesis by Pechini’s
method at the Biotecnology Department, University of Verona.
3.5 Nanoparticles coating

Recently, the fabrication of core-shell structures has attracted increasing attention. Ini-
tially, the idea of surface modification was applied in semiconductor nanoparticles but it
has also been extended to inorganic luminescent systems [18], [19], [20]. The formation
of core-shell structures is an effective way to improve luminescence efficiency since the
shell protects the core from interaction with the surrounding medium [21]. Moreover, in
some cases such as SiO2 , the coating allows the nanocrystals to be redispersible in organic
solvents.
A system of chemical reactions to control the growth of spherical silica particles with
sizes in the range 50-200 nm was developed by Stöber in 1968 [22]. A procedure to
incorporate radioactive isotopes into silica particles was also established by Flachsbart and
Stöber [2]. This technique allows the control of the spherical shape for the nanoparticles,
and it was used for the SiO2 coating of Eu3+ doped YAG and Y2 O3 [20], [23].
In this work, SiO2 -coated nanoparticles were obtained from the above Y2 O3 , GGG
and YAG samples using the Stöber method [23], [2]. The experimental details were as
follows: a nanopowder sample (0.153 g) was mixed with 60 ml of isopropyl alcohol in a
100 ml flask. The nanoparticles were dispersed with ultrasonic oscillation for 45 minutes.
Then, 4.5 ml of deionized water and 6 ml of hydrous ammonia were gradually added under
stirring at 50 ◦ C for 10 minutes. After that, 145 µl of tetraethyl orthosilicate (TEOS) was
44 3.6. Precipitation method
added and the solution was stirred at 50 ◦ C for 1 hour. Finally, the reaction mixture was
centrifuged at 4000 rpm for 10 minutes. The resulting nanoparticles were washed three
times by centrifuging in deionized water at 4000 rpm for 10 minutes and dried at 90 ◦ C
for 20 hours.
3.6 Precipitation method

Different procedures have been reported for the synthesis of LMA, such as solid state
reactions [24], a Pechini sol-gel procedure [25], combustion synthesis [26] or a precipitation
method [27]. All these methods are known to produce LMA powders with grain size in
the range of micrometers. In this work, several attempts to prepare LMA host material
by intimately mixing the starting oxides, La2 O3 + 2MgO + 11Al2 O3 , and heating up
to 1400 ◦ C were performed, but a mix of the initial oxides or an amorphous phase was
obtained. Another problem related to the synthesis of this system was the presence of
Al2 O3 as a non-desired impurity. Combustion synthesis was supposed to give rise to single
phase LMA since it requires lower temperatures. However, our experience is that after a
combustion reaction and a heat treatment at 500 ◦ C, an amorphous phase was achieved.
Firing the samples at 1200 ◦ C for more than ten hours was needed in order to obtain
crystalline LMA by combustion. LMA synthesis following the precipitation method was
also performed.
Both methods, combustion and precipitation, led to the formation of crystalline LMA
with traces of Al2 O3 impurities; but, as we will see later (see Section 5.3.1), luminescence
properties are much more interesting for precipitation samples. Therefore, the synthe-
sis of microcrystalline powders of LMA co-doped with different concentrations of Mn2+
and Yb3+ was developed in this work following the precipitation method described in
Ref. [27]. Stoichiometric quantities of La(NO3 )3 5H2 O, Al(NO3 )3 9H2 O, Mg(NO3 )2 6H2 O,
Mn(NO3 )2 H2 O and Yb(NO3 )3 5H2 O were dissolved in distilled water under stirring. Then,
NH4 OH was added dropwise until a pH of 8.5 was reached and precipitation as hydroxides
occurred. The obtained hydroxides were centrifuged at 3000 rpm for 5 minutes, washed
with distilled water, and heated at 90 ◦ C for 20 hours. A treatment at 700 ◦ C for 2 hours
and 1500 ◦ C for 1 hour was carried out to obtain the material in its final form.
3. Synthesis 45
3.7 Colloidal semiconductor nanocrystals
The optical and magnetic properties of wide-gap diluted magnetic semiconductors (DMSs)
are attracting broad interest in both fundamental and applied research areas [28]. Two
different methods were used for the synthesis of colloidal W-ZnO nanoparticles doped
with TM ions such as Co2+ or Mn2+ , and W-CoO nanocrystals.
Pure ZnO and Zn1−x TMx O with x <30% for Co and x <2% for Mn QDs were prepared
using hydrolysis and condensation of acetates ((OAc)2 ) solution in dimethyl sulfoxide
(DMSO) [29], [30]. The experimental procedure was as follows: Acetates were dissolved
in DMSO, typically 9·10−3 moles of Zn(OAc)2 and Co(OAc)2 or Mn(OAc)2 in 90 ml
of DMSO. Simultaneously, tetramethyl ammonium hydroxide (N(Me)4 OH) (1.8 equiv
of OH− ) was dissolved in 30 ml of ethanol. Then, the N(Me)4 OH solution was added
dropwise to the (OAc)2 in DMSO under constant stirring and the reaction took place.
After a rapid nucleation, the reaction Ostwald ripens, which means that small crystals
are unstable and shrink due to their large surface-to-volume ratio while larger particles
are stable and grow [31]. The reaction time and the addition of N(Me)4 OH are important
factors in this synthesis. Bigger nanoparticles were obtained for longer reaction times.
When Co(OAc)2 or Mn(OAc)2 concentration increases, more time is needed, not only for
the reaction to occur, but also for the nanoparticles to grow. Dopants are excluded from
the initial nucleation but are incorporated into the ZnO nanoparticles almost isotropically
during growth from solution. The formation of ZnO becomes linear with base addition
above 1.0 equiv of OH− . Figure 3.4. shows the sequence of a reaction flask during the
synthesis of a Co2+ -doped ZnO sample. The first image, before the base addition, shows
the characteristic pink color of octahedral Co2+ in (OAc)2 . N(Me)4 OH addition turns
the solution from pink to blue, indicating the tetrahedral coordination of Co2+ in ZnO.
Nanocrystals prepared by this method were precipitated in ethyl acetate, washed with
ethanol and precipitated again by adding heptane. Colloidal nanocrystals were usually
stabilized by a layer of surfactants attached to their surface. The energy with which
surfactant molecules adhere to the surfaces of growing nanocrystals is one of the most
important parameters influencing crystal growth. Murray et al. discussed this fact for
46 3.7. Colloidal semiconductor nanocrystals
the growth of CdSe nanoparticles capped with trioctylphosphine oxide (TOPO) [32]. For
the final capping treatment, melted TOPO was added to the nanocrystals and heated at
180 ◦ C for 30 minutes. The resulting powders were precipitated in ethanol or methanol
and resuspended in toluene.
Figure 3.4: Preparation of ZnO: Co2+ colloidal nanocrystals at the Chemistry Department, University
of Washington. The solution changes from pink to blue indicating the conversion of octahedral Co2+ to
tetrahedral Co2+ .
Colloidal W-CoO nanoparticles were synthesized using a nonaqueous solution method

with the reaction system composed by the metal fatty acid salt, the corresponding fatty
acid and a hydrocarbon solvent [33], [30]. For preparing CoO nanoparticles, cobalt
stearate (Co(St)2 ) (0.625 g) was used as the precursor, a certain amount of stearic acid
(0.305 g) was employed as ligands and octadecene (ODE) (5 g) was chosen as the non-
coordinating solvent. Co(St)2 and stearic acid were dissolved in ODE stirring at 120 ◦ C
under vacuum. When the purple mixture was fully dissolved, it was placed in a N2 flow,
and it was rapidly heated to 320 ◦ C and held there under stirring for 1 hour. When the
solution turned greenish W-CoO nanocrystals were formed. After cooling to 100 ◦ C, the
obtained CoO particles were precipitated with ethanol, capped with TOPO, precipitated
in acetone and suspended in toluene.
Bibliography
[1] M.P. Pechini. U.S.A. Patent 3.330.697. 1967.
[2] H. Flachsbart and W. Stöber. Preparation of Radioactively Labeled Monodisperse

Silica Spheres of Colloidal Size. J. Colloid Interface Sci., 30: 568–573, 1969.
[3] C.C. Koch. Synthesis of nanostructured materials by mechanical milling: Problems

and oportunities. Nanostruct. Mater., 9: 13–22, 1997.
[4] S.R. Mishra, G.J. Long, F. Grandjean, R.P. Hermann, S. Roy, N. Ali, and A. Viano.
Magnetic properties of iron nitride-alumina nanocomposite materials prepared by
high-energy ball milling. Eur. Phys. J. D, 24: 93–96, 2003.
[5] V.V. Boldyrev. Mechanochemistry and mechanical activation of solids. Solid State
Ionics, 63-65: 537–543, 1993.
[6] S. Mørup, J.Z. Jiang, F. Bødker, and A. Horsewell. Crystal growth and the steady-
state grain size during high-energy ball-milling. Europhys. Lett., 56: 441–446, 2001.
[7] R.E. Thoma, G.M. Hebert, H. Insley, and C.F. Weaver. Phase equilibria in the
System Sodium Fluoride-Yttrium Fluoride. Inorg. Chem., 2: 1005–1012, 1963.
[8] N. Martin, P. Boutinaud, R. Mahiou, J.C. Cousseins, and M. Bouderbala. Prepa-

ration of fluorides at 80 ◦ C in the NaF(Y, Yb, Pr)F3 system. J. Mater. Chem,
9: 125–128, 1999.
[9] J. Lu, Q. Zhang, and F. Saito. Mechanochemical Synthesis of Nano-sized Complex

Fluorides from Pair of Different Constituent Fluoride Compounds. Chem. Lett.,
1: 1176–1177, 2002.
47
48 BIBLIOGRAPHY
[10] T. Tsuzuki and P.G. McCormick. Mechanochemical synthesis of metal sulphide

nanoparticles. Nanostruct. Mater., 12: 75–78, 1999.
[11] T. Tsuzuki and P.G. McCormick. Synthesis of CdS quantum dots by mechanochem-
ical reaction. Appl. Phys. A, 65: 607–609, 1997.
[12] T. Tsuzuki, J. Ding, and P.G. McCormick. Mechanochemical synthesis of ultrafine

zinc sulfide particles. Physica B, 239: 378–387, 1997.
[13] J.A. Capobianco, F. Vetrone, J.C. Boyer, A. Speghini, and M. Bettinelli. Enhance-
ment of Red Emission via Upconversion in Bulk and Nanocrystalline Cubic Y2 O3 :
Er3+ . J. Phys. Chem. B, 106: 1181–1187, 2002.
[14] J.A. Capobianco, F. Vetrone, T. D’Alessio, G. Tessari, A. Speghini, and M. Bettinelli.

Optical spectroscopy of nanocrystalline cubic Y2 O3 : Er obtained by combustion
synthesis. Phys. Chem. Chem. Phys., 2: 3203–3207, 2000.
[15] T. Ye, Z. Guiwen, Z. Weiping, and X. Shangda. Combustion synthesis and photolu-
minescence of nanocrystalline Y2 O3 : Er phosphors. Mater. Res. Bull., 32: 501–506,
1997.
[16] M. Daldosso, D. Falcomer, A. Speghini, P. Ghigna, and M. Bettinelli. Synthesis,

EXAFS investigation and optical spectroscopy of nanocrystalline Gd3 Ga5 O12 doped
with Ln3+ ions (Ln=Eu, Pr). Opt. Mat., 30: 1162–1167, 2008.
[17] A.V. Rosario and E.C. Pereira. The effect of composition variables on precursor
degradation and their consequence on Nb2 O5 film properties prepared by the Pechini
Method. J. Sol-Gel Sci. Techn., 38: 233–240, 2006.
[18] J.S. Steckel, J.P. Zimmer, S. Coe-Sullivan, N.E. Stott, V. Bulovic, and M.G.
Bawendi. Blue Luminescence from (CdS)ZnS Core-Shell Nanocrystals. Angew.
Chem., 116: 2206–2210, 2004.
[19] O. Lehmann, K. Kömpe, and M. Haase. Synthesis of Eu3+ -Doped Core and
Core/Shell Nanoparticles and Direct Spectroscopic Identification of Dopant Sites at
BIBLIOGRAPHY 49
the Surface and in the Interior of the Particles. J. Am. Chem. Soc., 126: 14935–14942,
2004.
[20] D. Hreniak, P. Psuja, W. Strȩk, and J. Hölsä. Luminescence properties of Y3 Al5 O12 :
Eu3+ -coated submicron SiO2 particles. J. Non-Cryst. Solids, 354: 445–450, 2008.
[21] P. Zhu, Q. Zhu, H. Zhu, H. Zhao, B. Chen, Y. Zhang, X. Wang, and W. Di. Effect
of SiO2 coating on photoluminescence and thermal stability of BaMgAl10 O17 : Eu2+
under VUV and UV excitation. Opt. Mat., 30: 930–934, 2008.
[22] W. Stöber, A. Fink, and E. Bohn. Controlled Growth of Monodisperse Silica Spheres
in the Micron Size Range. J. Colloid Interface Sci., 26: 62–69, 1968.
[23] Q. Lü, A. Li, F. Guo, L. Sun, and L. Zhao. The two-photon excitation of SiO2 -
coated Y2 O3 : Eu3+ nanoparticles by a near-infrared femtosecond laser. Nanotech.,
19: 205704–205712, 2008.
[24] W. Ge, H. Zhang, J. Wang, D. Ran, S. Sun, H. Xia, J. Liu, X. Xu, X. Hu, and
M. Jiang. Growth and thermal properties of Co2+ : LaMgAl11 O19 crystal. J. Crystal
Growth, 282: 320–329, 2005.
[25] P.Y. Jia, M. Yu, and J. Lin. Sol−gel deposition and luminescent properties of
LaMgAl11 O19 : Ce3+ /Tb3+ phosphor films. J. Solid State Chem., 178: 2734–2740,
2005.
[26] S. Ekambaram, K.C. Patil, and M. Maaza. Synthesis of lamp phosphors: facile
combustion approach. J. Alloys Compd., 393: 81–92, 2005.
[27] J.M.P.J. Verstegen, J.L. Sommerdijk, and J.G. Verriet. Cerium and Terbium lumi-
nescence in, LaMgAl11 O19 . J. Lumin., 6: 425–431, 1973.
[28] J.M.D. Coey. Dilute magnetic oxides. Curr. Opin. Solid State Mater. Sci., 10: 83–92,
2006.
50 BIBLIOGRAPHY
[29] D.A. Schwartz, N.S. Norberg, Q.P. Nguyen, J.M. Parker, and D.R. Gamelin. Mag-
netic Quantum Dots: Synthesis, Spectroscopy, and Magnetism of Co2+ - and Ni2+ -
doped ZnO Nanocrystals. J. Am. Chem. Soc., 125: 13205–13218, 2003.
[30] M.A. White, S.T. Ochsenbein, and D.R. Gamelin. Colloidal Nanocrystals of Wurtzite
Zn1−x Cox O (0 6 x 6 1) : Models of Spinodal Decomposition in an Oxide Diluted
Magnetic Semiconductor. Chem. Mater., 20: 7107–7116, 2008.
[31] Y. Yin and A.P. Alivisatos. Colloidal nanocrystal synthesis and the organic-inorganic
interface. Nature, 437: 664–670, 2005.
[32] C.B. Murray, D.J. Norris, and M.G. Bawendi. J. Am. Chem. Soc., 115: 8706–8715,
1993.
[33] N.R. Jana, Y. Chen, and X. Peng. Size- and Shape-Controlled Magnetic (Cr, Mn,
Fe, Co, Ni) Oxide Nanocrystals via a Simple and General Approach. Chem. Mater.,
16: 3931–3935, 2004.
Chapter 4
Experimental Methods
4.1 Introduction
Along this chapter, we show the experimental procedures and equipment utilized not
only for the structural characterization of the samples, but also for the spectroscopic
study developed. First, XRD and transmission electron microscopy (TEM) techniques
are detailed. Then, the different setups used for the spectroscopic measurements are
described. Absorption, luminescence, excitation, lifetime and Raman measurements have
been performed. Besides, low temperature and high pressure techniques are described.
Finally, the spectra correction procedures are explained.
4.2 Structural characterization
Different techniques such as XRD or TEM were employed to characterize the prepared
materials. XRD measurements were accomplished in order to check the phase purity of all
samples and to estimate the particle size in the case of nanocrystalline samples. Average
particle size and size distribution of the nanoparticles were also estimated from TEM
images.
4.2.1 X-ray diffraction
XRD technique is commonly applied for the structural characterization of all kind of
materials. It is based on the diffraction produced when a X-ray beam is scattered by a
material. If the material presents crystalline structure, the scattered waves interfere in a
51
52 4.2. Structural characterization
constructive way for specific directions according to Bragg’s law, 2d sin θ = nλ, where d
is the distance between crystallographic planes, λ is the incident X-ray wavelength, θ is
the angle between the incident wave and the dispersion plane, and n is an integer number
indicating the diffraction order.
Various X-ray diffractometers were used to obtain XRD patterns (see Fig. 4.1). A
Philips 1700 diffractometer with a Bragg-Brentano θ − 2θ geometry, in which the X-ray
source is fixed, the sample turns θ and the detector turns 2θ, was utilized at the Solid State
Physics laboratory (University of Cantabria). A Thermo Electron ARL X’TRA based on
a vertical θ − θ geometry, where both the X-ray source and the detector move an angle
θ, was employed during a collaboration with Prof. Marco Bettinelli at the Biotecnology
Department (University of Verona). Both diffractometers are equipped with a Cu-anode
X-ray source (Kα , λ=1.5418 Å). The phase of the samples synthesized at Verona was
identified by comparing the obtained XRD patterns with the PDF 4+ 2006 database and
using SIeve program, while MAUD program was used for Rietveld data refinement [1].
Figure 4.1: X-Ray diffractometers employed to characterize our samples, Philips 1700 (left) and Thermo
Electron ARL X’TRA (right).
High pressure XRD experiments were developed with a Xcalibur diffractometer (Ox-
ford Diffraction Limited) (Fig. 4.2) at the Complutense University of Madrid within the
MALTA-Consolider project. XRD patterns were obtained on a 135 mm Atlas charge-
coupled device (CCD) detector placed at a distance of 90 mm from the sample using
Kα1 : Kα2 molybdenum radiation (λ=0.7107 Å). The X-ray beam was collimated to a
diameter of 300 µm. Exposure times were generally of 1 hour and the accessible angular
range is 4θ=60◦ . The observed intensities were integrated as a function of 2θ in order
to give conventional one-dimensional diffraction profiles. The CrysAlis software (Oxford
4. Experimental Methods 53
Diffraction Limited) was used for the data collection and the preliminary reduction of
the data. The indexing and refinement of the powder patterns were performed using the
TOPAS package.
Figure 4.2: Xcalibur diffractometer (Oxford Diffraction Limited) used for the high pressure measure-
ments (left). A CCD image of a CdS XRD pattern under pressure is also shown (right).
The particle size can be determined by X-ray peak broadening analysis for nanocrys-
talline samples with size up to 500 nm [2]. The peaks diffraction broadening arises mainly
due to three factors: Instrumental effects, crystalline size and lattice strains. The peak
width due to instrumental broadening is estimated using a standard Si sample. If the
observed XRD peak has a width B0 , and the width due to instrumental effects is Bi ,
then Br , which corresponds to XRD peak width at half-maximum once the instrumental
q p
broadening has been substracted, Br = (B0 − Bi ) B02 − Bi2 , is due to both crystalline
size and lattice strains. Broadening of XRD peaks due only to crystalline size is given by
kλ
Scherrer formula, Bc = L cos θ
, where λ is the X-ray wavelength, θ is the Bragg angle, L
is the average particle size and k is a constant. The broadening caused by lattice strains
can be represented by Bs = η tan θ, where η is the strain in the material. From these
expressions, the Williamson-Hall equation to determine the particle size is deduced [3]:
kλ kλ
Br = Bc + Bs = + η tan θ → Br cos θ = + η sin θ (4.1)
L cos θ L
When Br cos θ is plotted against sin θ a straight line is obtained; crystalline size can
kλ
be calculated from the intercept value L
[2]. The Williamson-Hall approach separates
the effect of size and strain in the nanocrystals. Spherical shape is assumed for the
nanoparticles. Strain contribution to the line broadening is significant for nanoparticles
grown by mechanical procedures.
4.2.2 Transmission electron microscopy
TEM is a technique of great importance as far as structural characterization of nanometric

systems is concerned. In TEM measurements, electrons are accelerated and focused on the
specimen (sample) by the condenser system, usually two lenses. After passing through the
sample, electrons are collected and the objective lenses form both images and diffraction
patterns. The main parts of the microscope are: Electron source, sample holder, light
and electron optics, electron detection and display. The TEM resolution, the smallest
0.61λ
distance that can be resolved δ, is given by δ = β
, where λ is the wavelength of the
radiation and depends on the electrons energy, and β is the semi-angle of collection of the
magnifying lenses. One important drawback of the TEM is that only a small part of the
sample is seen at a time. Moreover, TEM presents 2D images of 3D specimens, so the
information is averaged through its thickness. Another limitation is related to the fact
that samples must be ”electron transparent”, which means they must be thinner than
100 nm. For our powder specimen, preparation is easier than for thin films or metallic
samples, and involves selecting fine powder, suspending it in a non reactive liquid, and
placing it on the grid [4].
TEM experiments were performed on a JEM 2100 electron microscope (JEOL, Japan)
(Fig. 4.3) at the TEM facility at the University of Cantabria (SERMET), operating at
an accelerating voltage of 200 kV, with an estimated point to point resolution of 0.23
nm, and equipped with a CCD detector (Gatan Orius SC 1000B ). DigitalMicrograph
(Gatan) software was employed for the data collection and treatment. The samples were
prepared by suspending the solid powder in ethanol (Panreac, 96 purity) under ultrasonic
vibration. One drop of the prepared suspension was applied to carbon films on copper
grids (Agar Scientific).
Figure 4.3: JEM 2100 transmission electron microscope (JEOL, Japan) used at the TEM facility at
the University of Cantabria.
4.3 Spectroscopy
The aim of this work is the study of the optical properties of the synthesized materials.
Different spectroscopic techniques were used to analyze the involved transitions in the
light-matter interaction. As it is described below, absorption, luminescence, excitation,
time resolved spectroscopy, lifetime and Raman spectroscopy techniques were applied.
The spectroscopic measurements were carried out at the High Pressure and Spectroscopy
group laboratories at the University of Cantabria.
4.3.1 Absorption
A light beam becomes attenuated in a material according to Lambert-Beer law, I =

I0 exp(−αx), where I0 is the incoming intensity, I is the light intensity after passing
through a thickness x of the sample, and α is the absorption coefficient. Absorption
spectra involve the measurement of the light attenuated by a sample as a function of
the radiation wavelength, A(λ). The absorbance is defined as A = log II0 . In the case of
powder materials, absorption can only be measured by using very thin samples, although
reflectance spectra provide similar information. The spectral reflectance is defined as the
ratio of the flux reflected by the specimen to that of a standard surface under identical
56 4.3. Spectroscopy
geometrical and spectral conditions. Reflectance spectra can be registered in two different
modes: specular reflectance and diffuse reflectance, for the latter, an integrating sphere
is required to collect the diffuse reflected light for different wavelengths [5].
Figure 4.4: Cary 6000i spectrophotometer employed for reflectance measurements. The optical design
of the diffuse reflectance sphere (a) and the two components of the reflected light, specular and diffuse
reflection (b) are shown.
A Cary 6000i (Varian) spectrophotometer was used to register diffuse reflectance spec-
tra in the range 200-1800 nm (Fig. 4.4). It is equipped with two light sources; a quartz
halogen lamp for the visible/IR region and a deuterium lamp for the UV, and two de-
tectors; one photomultiplier (PMT) (Hamamatsu R928 ) for the visible region and one
InGaAs detector for the near IR. Diffuse reflectance measurements were performed using a
polytetrafluoroethylene (PTFE)-coated integrating sphere. Initially, a baseline is recorded
with the PTFE reference disk covering the reflectance port. The sample is then mounted
over the port and the light reflected by the sample surface is collected by the sphere.
The total (diffuse and specular) or the diffuse-only reflectance (see Fig. 4.4(b)) may be
measured by mounting the sample against the sphere port in two different configurations.
A FT-IR System 2000 (Perkin-Elmer) spectrophotometer was used to obtain the trans-
mission spectra in the IR region. It consists essentially of three components: Two radi-
ation sources to cover the 4400-400 cm−1 range, one Michelson interferometer and a
Mercury-Cadmium-Telluric detector. In addition, IR absorption spectra of the samples
in the form of KBr pellets were recorded on another spectrophotometer (Nicolet Magna
760 ), with a Deuterated-Triglycine-Sulfate detector.

Absorption experiments at high pressure requires non-conventional home-made se-
tups. The experimental apparatus used in this work is shown in Fig. 4.5. Tungsten
and deuterium lamps attached to a monochromator (Acton Research Corporation Spec-
tra Pro-300i) were used to obtain monochromatic light. One parabolic mirror and two
reflection objectives were incorporated in order to avoid chromatic aberration. The first
objective focuses the excitation beam in the hydrostatic cavity and the second one collects
the transmitted light, which is measured using a PMT with a lock-in technique (Stanford
Research systems Sr830 DSP ). A spatial filter must be used before the first objective.
Figure 4.5: Optical setup to measure high pressure absorption at the High Pressure and Spectroscopy
group laboratories, University of Cantabria.
4.3.2 Luminescence and excitation
In luminescence spectra, the emitted light intensity is measured as a function of the wave-
length for a fixed excitation wavelength, I(λ). On the contrary, in excitation technique,
only one wavelength is detected and the excitation wavelength is scanned in a certain
spectral range. Both luminescence and excitation, in contrast to absorption, are selective
techniques. Different light sources in the IR-visible-UV range, diverse detection systems
sensitive to those spectral regions and several monochromators were used to obtain this
kind of spectra. All spectra were corrected for the system response and were represented
as photons counts vs wavenumbers (see Section 4.6).
A standard spectrofluorimeter (Jobin-Yvon Fluorolog-2 ) was employed for the mea-

surement of photoluminescence and excitation spectra at room temperature (RT) and low
temperatures (Fig. 4.6). The sample is excited with a Xe-lamp followed by an excita-
tion monochromator, and the emitted light is detected with a PMT (Hamamatsu R928 )
at a 90◦ configuration in order to avoid direct light from the lamp. Continuous wave
(CW) UC luminescence spectra were obtained by exciting with a CW laser-diode (LD)
(LUMICS GmbH & Arroyo Instruments) and using the Fluorolog-2 detection system (see
setup scheme in Fig 4.6(a)). The LD can reach a maximum output power of 4 W and its
emission wavelength can be slightly tuned by changing the temperature. To obtain the
spectra, the nanopowders were transferred into a quartz capillary and closed after partial
air evacuation. A special sample holder was designed and adapted to the Fluorolog-2 (Fig.
4.6(b)). This allows to measure all the samples in identical conditions. Yb3+ luminescence
spectra were recorded by exciting with the LD in a single monochromator equipped with
an extended IR PMT (Hamamatsu R7102 ).
Figure 4.6: Jobin-Yvon Fluorolog-2 fluorimeter used for luminescence and excitation measurements.
The adapted setup for temperature-dependent UC experiments upon excitation with a LD (a) and the
sample holder (b) are also shown.
Another setup utilized to obtain luminescence spectra is shown in Fig. 4.7. An optical
parametric oscillator (OPO) laser system (Opotek Vibrant model B 355 II ) pumped by
the third harmonic of a Q-switched Nd:YAG (Brilliant Quantel) was used as excitation
source. This tunable OPO system offers an approximate pulse width of 10 ns with a
repetition rate of 10 Hz, and a continuous tuning range of 410-710 nm in the visible region
(signal mode) and 710-2400 in the IR region (idler mode). Pulses were guided through
a coupled optic fiber to the sample. The emitted light is collected by a 10× Mitutoyo
objective and guided through a fiber to the detection system. A monochromator (TRIAX
320) and an intensified CCD camera (Horiba-Jobin-Yvon iCCD 3553 ) were used for the
visible and UC luminescence intensity detection.
Figure 4.7: Optical setup used for visible and UC luminescence experiments at the High Pressure and
Spectroscopy group laboratories, University of Cantabria.
4.3.3 Lifetime and time resolved spectroscopy
To obtain information about both radiative and non-radiative decay from an excited state
and energy transfer processes, the luminescence temporal evolution after pulsed excitation
must be measured. This was performed in two different manners:
• Recording the temporal dependence of the luminescence intensity at a specific wave-

length. For the simplest case, in which excitation and emission occur in the same
center, an exponential decay is measured. Fitting this curve to a single exponen-
tial, I(t) = I0 exp( −t
τ
), the excited state lifetime, τ , is obtained. On the contrary,
when energy transfer processes are involved, an intensity rise followed by a decay is
detected once the excitation pulse has stopped.
• Obtaining the whole emission spectra at different times after the excitation pulse
has reached the sample. This experimental procedure is called time-resolved lumi-
nescence and the emission spectrum is recorded at a certain delay time with respect
to the excitation pulse and within a temporal gate. This technique is very useful to
explore energy transfer processes and allows us to identify contributions of differ-

ent optically active centers present in a sample if they have different desexcitation
dynamics.
Figure 4.8: Optical setup for lifetime measurements at the High Pressure and Spectroscopy group
laboratories, University of Cantabria.
For fluorescence lifetime experiments, the 10 ns laser pulses of the OPO system or
modulated LD excitation were used. In the case of this last device, the modulation
frequency must be suitable so that the system can reach the steady state and the lifetime
can be measured. The setup employed for lifetime measurements upon OPO pulsed
excitation is shown in Fig. 4.8. The sample luminescence was dispersed by a 0.50 m single
monochromator (CHROMEX 500IS/SM) equipped with 500 nm blazed 1200 grooves/mm
and 750 nm blazed 600 grooves/mm gratings, detected by a PMT (Hamamatsu R928 ) or
extended IR PMT (Hamamatsu R7102 ) and recorded with a multichannel scaler (Stanford
Research SR-430). An appropriate load resistance was placed between both devices. The
time constant of the RC circuit (C = 10 pF ) should be much lower than the luminescence
lifetime, τ , in order to avoid deconvolution processes. The smallest load resistance, R =
20 Ω was used in all measurements. The OPO+iCCD setup described in the previous
section (fig. 4.7) was also used for lifetime and time-resolved luminescence experiments.
4.3.4 Raman
Raman spectroscopy deals with the study of the fraction of incident light scattered by a
sample. Rayleigh scattering is an elastic photon process in which the scattered photon
energy is equal to the incident photon energy. However, in Raman spectroscopy the light
is inelastically scattered by a substance. The Stokes and anti-Stokes Raman scattering
involve virtual levels which do not correspond to real states, as a result, Raman spectra are
much weaker than fluorescence spectra by a factor of about 106 −108 . Raman experiments
are usually carried out under non-resonant illumination so that the Raman spectrum is
not masked by the more efficient emission spectrum [5], [6]. A typical Raman spectrum
shows an intense band at the incident frequency, ω0 , resulting from Rayleigh scattering,
and fainter Raman bands on both sides of ω0 at distances corresponding to vibrational
frequencies, ωk . In practice, the Stokes Raman scattering is more intense than the anti-
Stokes and the lower spectral region is used [7].
Figure 4.9: Equipment used for Raman measurements at the High Pressure and Spectroscopy group
The T64000 Raman spectrometer system (Horiba), together with a Kripton-Argon

laser (Coherent Innova Spectrum 70C) (the 514.5 nm green line was used), and a Nitro-
gen cooled CCD (Jobin-Yvon Symphony) with a confocal microscopy for the detection
(Fig. 4.9), were employed for Raman experiments. The T64000 system is composed of
three monochromators (640 mm focal length) with two basic configurations; triple addi-
tive mode, with the three monochromators operating in series (3×640 mm focal length),
62 4.4. Temperature dependence
and substractive mode, in which the two first monochromators filter the laser, and the
analysis is done with the third one. Different high pressure cells can be adapted to the
XYZ automated stage for high pressure measurements. LabSpec software permits data
acquisition and treatment.
4.4 Temperature dependence
Many of the spectroscopic characterization measurements, such as luminescence, excita-

tion or lifetime, were accomplished as a function of temperature.
• Low temperature measurements were achieved using a closed-cycle helium cryo-

stat (Air Products CS202E ) (Fig. 4.10). The major components are the expander
or cold finger, the compressor, the radiation shield and the vacuum shroud. The
sample is placed on the cold finger using Cry-Con copper-filled thermal grease to
fix the sample and to ensure thermal conductivity at low temperatures. This sys-
tem requires a vacuum pump for the sample space. A combination of a diffusion
pump (Leybold PD 180 L) with a rotatory pump (Leybold Trivac B ) as well as a
turbomolecular pump (Varian Turbo Dry 70 ) were employed in our experiments.
A programmable temperature controller (APD-K cryogenics HC-2 ) together with
a Si-diode thermocouple let us modify and stabilize the temperature in the range
10-300 K with an accuracy better than 0.1 K.
Figure 4.10: Closed-cycle helium cryostat components (left) and compressor (right) used for low tem-
perature experiments down to 15 K.
• High temperature experiments (300-650 K) were performed using a microscope heat-

ing stage (Leitz 350) (Fig. 4.11). The sample is placed between two quartz films and
put into a metallic cavity. The heat is controlled by a rheostat and a thermocouple
is used to measure the temperature.
Figure 4.11: Heating stage, power supply and thermometer for the temperature measurements used in
the range 300-650 K.
4.5 High pressure measurements

High pressure studies are a valuable mechanism to explore changes in materials as a
function of volume without changing the chemical composition. Hydrostatic pressure
is used to systematically influence the bonding environment of luminescent centers and
therefore modify electronic states energy and transitions probabilities. Sometimes, high
pressure can induce structural phase transitions and even stabilize structures not achiev-
able through other means. As a consequence, their optical properties change [8]. High
pressure is also a very efficient tool for understanding the electronic structure of semi-
conductors. The interatomic distance is reduced in average when hydrostatic pressure
is applied to a material, and this fact normally causes an increase in the semiconductor
bandgap energy [9].
4.5.1 Diamond anvil cells
High pressure measurements were carried out in diamond anvil cells (DACs). The use of
diamond has two important advantages over other anvils. Firstly, diamond is the hardest
known material and is capable of reaching higher pressures [10]. Secondly, diamond is
transparent in a wide frequency range including X-ray and IR-visible range.
64 4.5. High pressure measurements
Figure 4.12: Hydrostatic cavity in the cell obtained after placing the perfored gasket containing the
sample between the two opposite diamond anvils.
The basic principle of the DAC is very simple. A metal gasket containing the sample,
the hydrostatic medium and ruby chips is placed between the flat parallel faces of two
diamonds (see Fig. 4.12). By pushing the two opposed anvils together, the sample pressure
increases. The gasket is prepared by drilling a hole at the center of the indentation made
by the anvil face. Then, the sample, the medium, and a ruby chip if necessary, are placed
inside the cavity [11].
Figure 4.13: Spark eroder used for gaskets perforation. This system allows the utilization of wires
between 25 and 400 µm. A picture of a perfored gasket is also shown.
Sample chamber preparation.
The metal foil or gasket serves three purposes; it provides the high-pressure sample cham-
ber, it avoids direct contact between diamonds, and it gives lateral support to the conical
faces of the anvils. The gasket is first indented by an anvil face, this process compresses
and hardens the material. The initial thickness of the foil is about 300 µm whereas the
indented thickness varies from 50 to 200 µm. The deeper the indentation is, the higher
pressures can be reached. Gasket perforation was performed using a semi-automatic spark
erosion machine (Betsa MH 20M) with tungsten electrodes (Fig. 4.13). The hydrostatic
cavity diameter was usually between 150 and 300 µm.
DACs specifications.
Different cells were used in the high pressure experiments at the High Pressure and Spec-
troscopy group laboratories at the University of Cantabria, as well as at the MALTA
X-Ray Diffractometer at the Complutense University of Madrid. A Brillouin-Raman cell
(Diamond Optics Inc.) (Fig. 4.14) and a Membrane cell (developed at the Pierre and
Marie Curie University) (Fig. 4.15) were utilized for absorption and Raman measurements
on CdS and ZnO nanocrystals under high pressure. A cryoDAC-Mega cell (easyLab Tech-
nologies) was also employed for high pressure luminescence and lifetime experiments on
GGG: Cr3+ nanoparticles. High pressure XRD measurements on CdS nanocrystals were
performed in a modified Merrill-Bassett (MALTA cell).
• Brillouin-Raman cell This cell is based on the modification made by Mao and
Bell of the Merril-Basset cell, and is designed for reaching pressures as high as 40
GPa [11], [12], [13]. The main body is made of two 440c steel pieces perfectly face to
face in which the two plates are fit. Diamond anvils are placed in the middle of the
plates. Pressure is applied by turning four allen screws which pull the two plates,
and hence the diamonds, together. Type-IIa diamonds with very low luminescent
impurities concentration, and with a 0.6 mm culet diameter, were used for the anvil.
• Membrane cell The main difference between the membrane cell and the Brillouin-
Raman cell is that the former uses a metallic diaphragm or membrane to generate
the pressure instead of screws. This cell is composed of two steel parts which fit
perfectly one into the other. The membrane, a toroidal cavity in which pressurized
gas is introduced, is mounted on the upper part of the cell. When the membrane
is inflated, it pushes the upper cylinder against the other increasing the pressure
Figure 4.14: Diamond Optics Brillouin-Raman cell used for absorption and Raman measurements at
high pressures.
Figure 4.15: Membrane cell used for absorption and Raman measurements at high pressures.
in the hydrostatic cavity. This system allows a fine pressure control and has the
advantage that there is no need of moving the cell to change the pressure. The
utilized diamond anvils had 0.5 mm of culet size.
• cryoDAC-Mega cell This cell, made of CuBe alloy, is suitable for both optical and
XRD studies at cryogenic temperatures. The diamond anvils are mounted within
CuBe rings and mechanically fixed to their tungsten carbide support plates. The
DAC itself is held within two CuBe clamps, and one of them is fixed to the cryostat.
Pressure is applied by turning the four bolts on the CuBe blocks.
• MALTA cell. In the case of this modified Merrill-Bassett DAC, three screws pull
the two plates together [11]. The used diamond anvils had 1 mm of culet size.
4.5.2 Pressure calibration and transmitting media
The most commonly used technique to determine the pressure inside the DAC cavity is
based on the pressure shift of the R-lines (2 E → 4 A2 ) ruby luminescence. The RT R1 and
R2 shift rate is linear up to 20 GPa according to [14]:
ER1 (cm−1 ) = 14405 − 7.53 · P (GPa)
ER2 (cm−1 ) = 14434 − 7.53 · P (GPa) (4.2)
Ruby shows a sharp and efficient luminescence facilitating its detection. However,
there are times, like for instance in the case of GGG: Cr3+ , when ruby emission can hide
the sample luminescence, and a non-luminescent pressure sensor must be used. Spectral
shift of Raman lines in diamond have also been calibrated as a function of pressure
[15]. The Raman-active phonon energy is DR = 1332.26 cm−1 at ambient pressure (AP).
Pressure can be expressed as a function of the Raman peak position [16]:
DR (cm−1 ) = 1332.26 + 2.64 · P (GPa) (4.3)
The function of the transmitting medium is to ensure a homogeneous pressure dis-

tribution in the sample chamber. Pressure gradients and stress must be reduced, since
they can alter the physical state of the sample independently of any hydrostatic pressure
effects. Basic requirements for the hydrostatic medium are being transparent in the wave-
length range under study, and keeping the hydrostaticity within a wide pressure interval.
Different transmitting media were used. Paraffin oil is hydrostatic below 10 GPa and it
was chosen as a good candidate because it is inert. When higher pressures were required,
silicon oil (Dow Corning) was employed. A 4:1 mixture of ethanol: methanol provides
hydrostatic conditions at RT up to 10 GPa and it was also used.
4.6 Spectra correction

The correction of the experimental excitation and luminescence spectra is important when
precise information about spectral positions, bands shape and intensity is desired. Two
68 4.6. Spectra correction
types of corrections were applied in this work.
4.6.1 Wavelength correction
In the spectra shown in this work, the x axis is represented as wavenumbers, in cm−1
units. However, the raw data were obtained as a function of wavelength, λair (nm). To
accomplish this conversion, first, we need to calibrate the monochromator using a Hg
lamp. Then, the refractive index of air must be considered in order to get the vacuum
wavelength, λvac (nm), according to the following expression [17]:
1.2288 3.555 · 104

λvac = λair + 2.72643 · 10−4 · λair + + (4.4)
λair λ3air
Finally, the vacuum wavelength, λvac (nm), is converted into wavenumbers, E(cm−1 ):
E(cm−1 ) = 107 · λ−1

vac (nm) (4.5)
4.6.2 Intensity correction
• Excitation spectra. In order to correct the excitation spectra intensity, we must

bear in mind the different radiation source power as a function of the wavelength,
the response R(λ). The excitation spectra in the visible region were achieved using
a Xe-lamp in the Fluorolog-2. The lamp power is measured simultaneously while a
spectrum is being collected. Yb3+ excitation spectra were recorded exciting with the
OPO laser. The laser power is monitored using a beam splitter and a powermeter
(Ophir 3A-SH ). In all cases, the excitation spectra were corrected by dividing them
by the corresponding response.
• Emission spectra. When luminescence spectra are measured, the detector signal
is determined by the amount of light emitted by the sample at each wavelength, and
the efficiency with which that light is detected. The experimental spectra must be
corrected so that they can be interpreted only in terms of the intrinsic emission of
the sample. The experimental procedure described by Ejder was followed [18]. The
signal can be expressed in terms of I (energy per second) or J (photons per second)
for y axis, and either λ(nm) or E(cm−1 ) for x axis. In this work, all the spectra are
represented as photons vs wavenumbers, so the integrated intensity of an emission

band is proportional to the number of emitted photons per time unit. First, the
spectrum of a black body source (calibrated filament lamp) was measured at the
same experimental conditions, B(λ). Then, the lamp spectrum was divided by the
theoretical output of the lamp, J(E):
1
J(E) ∝ E 2 (4.6)
exp( kBET ) −1
where T is the black body temperature (2535 K), and the detection system response
was obtained, R:
B(E)
R= (4.7)
J(E)
To correct a luminescence spectrum initially expressed in J(λ), LS (nm), one must

multiply by λ2 (nm), LS ’(nm)=LS (nm)·λ2 (nm), convert x axis to wavenumbers,
LS (cm−1 ) and divide by the response curve, R. The corrected emission spectrum is
given by LSc :
LS(cm−1 )
LSc = (4.8)
R
Bibliography
[1] L. Luterotti and S. Gialanella. X-ray diffraction characterization of heavily deformed

metallic specimens. Acta. Mater., 46: 101–110, 1998.
[2] C. Suryanarayana and M.G. Norton, editors. X-Ray Diffraction: A Practical Ap-
proach. Plenum Press, New York, 1998.
[3] G.K Williamson and W.H Hall. X-ray line broadening from filed aluminium and
wolfram. Acta. Metall., 1: 22–31, 1953.
[4] D.B. Williams and C.B. Carter, editors. Transmission Electron Microscopy. Plenum
Press, New York, 1996.
[8] K.L. Bray, editor. High pressure probes of electronic structure and luminescence prop-
erties of transition metal and lanthanides systems, volume 213 of Topic in Current
Chemistry. Spring-Verlag, Berlin, 2001.
[9] T. Suski and W. Paul, editors. Semiconductors and Semimetals, volume 54 of High
Pressure in Semiconductor Physics I. Academic Press, San Diego, 1998.
71
72 BIBLIOGRAPHY
[10] I.V. Aleksandrov, A.F. Goncharov, A.N. Zisman, and S.M. Stishov. Diamond at high
pressures: Raman scattering of light, equation of state, and high pressure scale. Sov.
Phys. JETP, 66: 384–390, 1987.
[11] A. Jayaraman. Diamond anvil cell and high-pressure physical investigations. Reviews
of Modern Physics, 55: 65–108, 1983.
[12] W.B. Holzapfel and N.S. Isaacs, editors. High-pressure techniques in Chemistry and
Physics. Oxford University Press, New York, 1997.
[13] W.F. Sherman and A.A. Stadtmuller, editors. Experimental techniques in high-
pressure research. John Wiley and Sons, New York, 1987.
[14] G.J. Piermarini, S. Block, J.D. Barnett, and R.A. Forman. Calibration of the pressure
dependence of the R1 ruby fluorescence line to 195 kbar. J. Appl. Phys., 46: 2774–
2780, 1975.
[15] D. Schiferl, M. Nicol, J.M. Zaug, S.K. Sharma, T.F. Cooney, S.Y. Wang, T.R. An-
13 12
thony, and J.F. Fleischer. The diamond C / C isotope raman pressure sensor
system for high-temperature/pressure diamond-anvil cells with reactive samples. J.
Appl. Phys., 82: 3256–3265, 1997.
[16] A. Tardieu, F. Cansell, and J.P. Petitet. Pressure and temperature dependence of
the first-order Raman mode of diamond. J. Appl. Phys., 68: 3243–3245, 1990.
[17] D.R. Lide, editor. Handbook of Chemistry and Physics. 65th edition. CRC Press,
1983.
[18] E. Ejder. Methods of Representing Emission, Excitation and Photoconductivity

Spectra. J. Opt. Soc. Am., 59: 223–224, 1969.
Chapter 5
Insulating materials. Luminescent

properties
5.1 Introduction
In this chapter, we present the most relevant results in relation with the electronic tran-
sitions responsible for the optical properties of different systems doped with RE and TM
ions. The synthesis of these materials has as its main goal the development of new lumines-
cent materials, specially those showing UC emission, and the study of the fundamental
physics, in particular the involved UC mechanisms, responsible for this luminescence.
First, luminescence properties of Y2 O3 and NaYF4 nanoparticles doped with Er3+ and
Yb3+ are investigated. The Er3+ -Yb3+ combination is frequently used in the development
of high efficiency UC materials [1], besides, NaYF4 is the most efficient host material up
to date for UC phosphors [2]. The underlying mechanisms in UC processes for Tb3+ -Yb3+
and Eu3+ -Yb3+ co-doped GGG and YAG nanocrystals are also analyzed. Then, green
UC Mn2+ luminescence in Mn2+ -Yb3+ co-doped LMA powders is demonstrated, as far as
we know, for the first time up to 650 K. In the last section, the temperature and pressure
dependence of the Cr3+ transitions in GGG nanoparticles is studied.
5.2 Rare-earth ions doped nanoparticles
UC materials, in which visible light is generated upon IR excitation, have attracted sig-
nificant attention for advanced applications such as solid state lasers [3], [4], materials
73
74 5.2. Rare-earth ions doped nanoparticles
for biological applications like bio-labeling, drug delivery, or diagnostics [5], [6], UC phos-
phors, IR quantum counter detectors, or efficiency improvement of bifacial solar cells [7].
UC laser systems show several advantages over direct UV excitation; IR excitation re-
duces the photo-ionization induced degradation of hosts, it does not need high excitation
wavelength stability, and the output wavelength is not restricted to a given harmonic [3].
Recently, Prasad et al. used upconverting nanophosphors for in vitro and in vivo pho-
toluminescence bio-imaging, providing deeper light penetration into the biological tissues
which are transparent in the 750-1000 nm range. The advantage of NIR excitation is the
reduction of the background autofluorescence of the tissues [6].
The vast majority of UC studies investigated up to date involve insulating materials

doped with RE ions. The energy emission in semiconductors is known to be very sensitive
to the system dimensionality and to the particle size [8]. However, due to the localized
character of the f electrons, the emission energies associated to f−f transitions in RE ions
are mainly independent of the nanocrystals size. On the other hand, the emission intensity
and lifetime may change when particle size decreases due to the presence of organic
impurities on the nanoparticles surface. The emission intensity can also be modified
by the combination of RE ions and by changing the doping concentrations. It is well-
known that Yb3+ is an excellent UC sensitizer for Er3+ , Tm3+ , Pr3+ or Ho3+ ions [1], [9].
All these ions present intermediate states almost resonant with the 2 F7/2 → 2 F5/2 Yb3+
transition. The doubly-doped UC systems containing Er3+ and Yb3+ present the highest
UC efficiencies [2]. On the contrary, the situation is completely different for RE ions like
Tb3+ or Eu3+ which have no intermediate levels resonant with Yb3+ . Tb3+ and Eu3+ are
attractive as emitting ions because of their high quantum efficiency related to the large
energy gap between the emitting states and the low lying 7 FJ (J = 0, 1, . . . , 7) excited
states.
The ability of Yb3+ to induce UC is based on the high oscillator strength of the unique
f−f transition, 2 F7/2 → 2 F5/2 , which is located in the NIR, just in the range of cheap and
high power LD. Yb3+ has no higher excited states and it is transparent in the visible
region. The UC processes involving resonant ions, like in Er3+ -Yb3+ co-doped materials,
can be ascribed to GSA/ESA and/or GSA/ETU mechanisms. Since the Tb3+ -Yb3+ UC
5. Insulating materials. Luminescent properties 75
system was introduced in 1969, many studies have been devoted to these two ions [10], [11],
[12], but the fact of getting UC emission in nanoparticles is not so common and always
noticeable. However, as far as we know, there are only few examples of UC luminescence
in Eu3+ -Yb3+ systems in the literature [13], [14]. The UC luminescence in Tb3+ -Yb3+
or Eu3+ -Yb3+ systems is assigned to a cooperative sensitization mechanism or GSA/ESA
processes in dimers.
In this work, nanocrystals of Y2 O3 : Er3+ , Yb3+ , NaYF4 : Er3+ , Yb3+ as well as
Tb3+ -Yb3+ and Eu3+ -Yb3+ co-doped GGG and YAG nanoparticles have been prepared
by using different synthesis methods. The potential applications of these nanocrystalline
materials go through the possibility of reaching the same properties observed in bulk but
in systems with dimensions in the range of nanometers. Silica is highly bio-compatible and
its surface chemistry is well documented for biological interactions; in this sense, coating
with SiO2 would be the first step in the functionalization of the obtained nanoparticles.
XRD, TEM, as well as RT luminescence, excitation and lifetime measurements have been
carried out in these samples. The most relevant structural and spectroscopic results in
order to characterize them and to identify the UC mechanisms in each case are shown in
this section.

Synthesis and characterization
Within inorganic materials, Y2 O3 is one of the most widely studied systems for accept-
ing trivalent impurities. It crystallizes in the cubic system, showing a bixbyte structure
((Fe,Mn)2 O3 ) with Ia3 space group (Fig. 5.1). The unit cell contains 16 molecules and
the bulk cell parameter is a0 =10.604 Å [15]. Y3+ ions are accommodated in 32 sites in the
unit cell; 24 sites with point group symmetry C2 and 8 sites with C3i symmetry. RE ions
substitute Y3+ ions and have been found to be randomly distributed in both sites [16].
The C3i site has associated a center of inversion, therefore, f-f electric dipole transitions
are much weaker than those related to the C2 site. The maximum phonon energy in bulk
yttria is 600 cm−1 [17].
In this work, Y2 O3 : Er3+ , Yb3+ nanoparticles prepared by two methods, ball milling
Figure 5.1: Y2 O3 cubic structure.
Figure 5.2: XRD patterns of Y2 O3 : 2%Er3+ , 1%Yb3+ nanoparticles prepared by ball milling for 2 and
180 hours (a), and calculated pattern for cubic Y2 O3 (a0 =10.604 Å) (b).
(top-down) and combustion (bottom-up) are studied. Both synthesis methods produce
nanoparticles of Y2 O3 cubic phase as it can be seen from XRD pattern. Figure 5.2 shows
the XRD patterns of the Y2 O3 : 2%Er3+ , 1%Yb3+ nanocrystalline samples prepared in the
planetary ball mill for 2 and 180 hours milling time according to the parameters defined in
Section 3.2. Considering the most intense peak in the XRD diagram for different grinding
times (Fig. 5.3), it can be observed that when grinding time increases, the peaks intensity
diminishes while their width increases. This broadening is due to both smaller particle
size and strains generated during the milling.
According to the Williamson-Hall equation (eq. 4.1), the average crystallite size can
Figure 5.3: Y2 O3 : 2%Er3+ , 1%Yb3+ XRD peak at 2θ=29.3◦ for different grinding times.
Figure 5.4: Average crystal size, L, for the Y2 O3 : 2%Er3+ , 1%Yb3+ ball milling sample as a function
of milling time.
be estimated from XRD diagrams. Considering the dependence of particle size on time
(Fig. 5.4), we realize that for times longer than 35 hours a steady-state particle size of
25 nm is reached. This limit value is a characteristic feature of each kind of material and
grinding process. We can conclude then, that the optimal milling time for doped Y2 O3
samples is around 50 hours.
Er3+ and Yb3+ co-doped Y2 O3 nanoparticles have also been prepared following the
glycine-nitrate solution combustion synthesis (reaction 3.4). Figure 5.5 shows the XRD
patterns obtained for a sample prepared by this method before and after it was treated at
500 ◦ C. It can be seen from the peaks broadening that in both cases the particle size is in
Figure 5.5: XRD patterns of Y2 O3 : 2%Er3+ , 1%Yb3+ nanocrystals prepared by combustion before
and after firing at 500 ◦ C (a), and calculated pattern for cubic Y2 O3 (Ia3 space group) (b).
Figure 5.6: Fitting of the experimental peak broadening to the Williamson-Hall equation for Y2 O3 :
2%Er3+ , 1%Yb3+ nanoparticles prepared by combustion and fired at 500 ◦ C for 1 hour.
the nanometer range. The fitting of the experimental peak broadening to the Williamson-
Hall equation for Y2 O3 : 2%Er3+ , 1%Yb3+ samples prepared by combustion and calcined
(Fig. 5.6) is given by Br cos θ = 0.01053 + 0.01407 sin θ. The average particle size of the
nanoparticles obtained by combustion is about 10 nm before calcination and it increases
to 15 nm after firing at 500 ◦ C for 1 hour.
The presence of CO2− −
3 and OH anionic groups on the nanoparticles surface is known
to be one of the main reasons for the diminishing of the emission intensity as well as for
the reduction of the emission lifetime. These ions contribute with high energy phonons
increasing the non-radiative relaxation processes via multiphonon relaxation according to
Figure 5.7: IR spectra of nanocrystalline Y2 O3 : 2%Er3+ , 1%Yb3+ prepared by combustion (up) and
ball milling (down). Sequential heat treatments have been carried out on combustion samples: 500 ◦ C
for 2 hours, and 1000 ◦ C for 5 h.
the gap’s law [17]. Moreover, this fact reveals surface effects in the luminescence properties
of nanocrystalline materials. The adsorption of these anionic groups is related to the
precursors used in the combustion synthesis but this cannot be the only reason, because
they also appear in the samples obtained by ball milling, where the starting materials are
free of these ions. In order to reduce surface contamination, different thermal treatments
have been performed on the nanocrystalline samples.
Figure 5.7 shows the medium infrared (MIR) spectra of Y2 O3 : 2%Er3+ , 1%Yb3+
prepared by both combustion and ball milling. All spectra show bands at approximately
1500 and 3400 cm−1 indicating the presence of CO2− −
3 and OH ions on the nanoparticles
surface. The intensity of these bands is reduced as calcination time and temperature
increase. The best nanoparticles were obtained after calcination at 1000 ◦ C for 5 hours.
Nevertheless, several hours after the heat treatment, the nanocrystalline samples were
able to adsorb again these anions.
Samples must be fired in order to reduce surface contamination but longer thermal
treatments at higher temperature sinterize the powders forming larger nanoparticles. Fig-
ure 5.8 presents the particle size dependence on calcination temperature, TC , and time,
tC , for Y2 O3 : 2%Er3+ , 1%Yb3+ combustion sample. It is observed that nanocrystals size
increases for longer firing times and higher temperatures.
Figure 5.8: Variation of Y2 O3 : 2%Er3+ , 1%Yb3+ nanoparticles size, L, as a function of the calcination
temperature (a), and time (b).
Figure 5.9: TEM image of Y2 O3 : 2%Er3+ , 1%Yb3+ prepared by combustion after SiO2 coating. In
the HRTEM images (b) and (c), the observed interplanar distances (0.33 nm) corresponds to the (222)
crystal plane.
Bearing in mind both effects, surface contamination and particle size, all Y2 O3 samples
studied in this work were fired at 900 ◦ C for an hour, and closed in shielded capillaries.
This covering was performed in order to avoid further contamination which produces a
deterioration of their optical properties. Some of the samples were also SiO2 coated. A
final average diameter of 50 nm has been estimated from the peak width of the XRD (see
Fig. 5.8). This value is in very good agreement with TEM results shown in figure 5.9.
TEM images of coated Y2 O3 nanoparticles show a SiO2 shell thickness of about 5-10 nm.
Optical properties
Optical properties of Y2 O3 : 2%Er3+ , 1%Yb3+ and Y2 O3 : 2%Er3+ , 20%Yb3+ nanoparticles

prepared by high energy ball milling and combustion synthesis are analyzed.
• Synthesis method and concentration dependence.
Figure 5.10 shows the RT diffuse reflectance spectrum of Y2 O3 : 2%Er3+ , 20%Yb3+

prepared by ball milling. The absorption peaks correspond to Er3+ transitions from
the ground state, 4 I15/2 , to different excited states, and to the 2 F7/2 → 2 F5/2 Yb3+
transition.
Figure 5.10: RT diffuse reflectance spectrum of Y2 O3 : 2%Er3+ , 20%Yb3+ prepared by ball milling.
Figures 5.11 and 5.12 compare the RT luminescence spectra of Y2 O3 doped with
different Er3+ and Yb3+ concentrations, and prepared by ball milling and combus-
tion, upon direct Er3+ excitation at 26596 cm−1 , and IR excitation at 10256 cm−1 .
Figure 5.11: RT emission of nanocrystalline Y2 O3 co-doped with different concentrations of Er3+

and Yb3+ upon excitation at 26596 cm−1 . Measurements were performed under the same experimental
conditions.
Figure 5.12: RT UC luminescence of Y2 O3 : Er3+ , Yb3+ nanoparticles with different RE concentrations

upon excitation at 10256 cm−1 . Measurements were performed under the same experimental conditions.
Figure 5.13: RT UC luminescence of Y2 O3 : 2%Er3+ , 20%Yb3+ nanoparticles prepared by combustion,

upon excitation at 10256 cm−1 , with and without SiO2 coating.
Emission peaks around 15000 and 18000 cm−1 are assigned, respectively, to the
transitions from the 4 F9/2 and the thermalized (2 H11/2 , 4 S3/2 ) excited states, to the
4
I15/2 ground state of Er3+ ions (Fig. 5.14). Figure 5.13 shows the UC emission
spectra of Y2 O3 : 2%Er3+ , 20%Yb3+ nanoparticles prepared by combustion with
and without SiO2 . The total intensity as well as the intensity ratio between red and
green emissions in the spectra of the nanoparticles with and without SiO2 coating
are exactly the same.
From the spectra in Fig. 5.11, we conclude that samples prepared by ball milling
have higher luminescence intensities than combustion samples. This is due to the
presence of CO2− −
3 and OH groups which is more important in combustion samples,
even after thermal treatment. A salient feature is that in the 2%Er3+ , 1%Yb3+
samples, the intensity of the green emission is higher than the red one. On the
contrary, those samples doped with 20%Yb3+ have a higher red to green intensity
ratio. This reduction in the green luminescence is related to the (2 H11/2 , 4 S3/2 ) +
2
F7/2 → 4 I11/2 + 2 F5/2 CR process between Er3+ and Yb3+ ions, whose probability
increases with increasing Yb3+ concentration, and is responsible for the additional
(2 H11/2 ,4 S3/2 ) depopulation channel (Fig. 5.14).
Figure 5.12 shows the UC luminescence spectra obtained at RT for Y2 O3 : 2%Er3+ ,

Figure 5.14: Energy level scheme of Er3+ and Yb3+ ions and the proposed UC mechanisms. Green
and red luminescence in Y2 O3 : Er3+ , Yb3+ samples are depicted. Cross relaxation is responsible for the
reduction of green emission. GSA/ESA or GSA/ETU from 4 I13/2 are responsible for the enhanced red
emission upon IR excitation.
1%Yb3+ and Y2 O3 : 2%Er3+ , 20%Yb3+ prepared by mechano-chemical and combus-

tion synthesis upon IR LD excitation centered at 10256 cm−1 . The broad excitation
(∆λ ∼ 10 nm) is able to excite both 2 F7/2 → 2 F5/2 Yb3+ and 4 I15/2 → 4 I11/2 Er3+
transitions. The same bands around 15000 and 18000 cm−1 are observed, and ball
milling nanocrystals are once again more efficient than combustion ones. The red UC
emission, 4 F9/2 → 4 I15/2 , is more intense than the green UC luminescence, (2 H11/2 ,
4
S3/2 ) → 4 I15/2 in all cases. For 20%Yb3+ -doped nanoparticles the green emission
is not detected by the naked eye. The enhancement of red to green emission upon
IR excitation can be explained with the following process: after IR excitation, both
2
F5/2 Yb3+ and 4 I11/2 Er3+ multiplets are populated. Yb3+ excitation is transferred
to the 4 I11/2 Er3+ level by energy transfer (GSA/ETU). Then, most of the ions decay
non-radiatively to the 4 I13/2 level, and Er3+ ions can reach the 4 F9/2 state after the
absorption of a second IR photon, ESA, or after another ETU process (Fig. 5.14).
In the absence of the 4 I11/2 → 4 I13/2 non-radiative process, a second photon could
reach the (2 H11/2 , 4 S3/2 ) green emitting states. This additional mechanism which
seems to be more efficient than the (2 H11/2 , 4 S3/2 ) → 4 F9/2 decay, does not exist
upon visible excitation.
• Temporal evolution and UC mechanism.
Temporal evolution of the red, 4 F9/2 → 4 I15/2 , and green, (2 H11/2 , 4 S3/2 ) → 4 I15/2 ,
Er3+ luminescence after pulsed excitation with the OPO at 20492 cm−1 and 10256
cm−1 has been recorded.
Figure 5.15 shows the time dependence of the red and green Er3+ luminescence of
Y2 O3 : 2%Er3+ , 20%Yb3+ prepared in the planetary ball mill upon excitation at
20492 cm−1 . Experimental data obtained after visible excitation have been fitted
to a single exponential, I(t) = I0 exp( −t
τ
) for all samples. The best fit parameters
are shown in Table 5.1. Both emitting states, 4 S3/2 and 4 F9/2 , show longer lifetimes
in ball milling samples in comparison with combustion ones. This fact reveals that
the intensity diminishing observed for samples prepared by combustion is related
to non-radiative processes, which are more important in samples prepared following
the combustion reaction.
Figure 5.15: Temporal evolution of the RT red (a) and green (b) Er3+ luminescence of Y2 O3 : 2%Er3+ ,
20%Yb3+ prepared by ball milling upon excitation at 20492 cm−1 .
Figure 5.16 compares the temporal behavior of the red Er3+ UC luminescence after
IR excitation at 10256 cm−1 for Y2 O3 : 2%Er3+ , 20%Yb3+ prepared by ball milling
and combustion synthesis. A rise of the emission intensity followed by a decay is
detected in both cases. In this case, the temporal evolution of the Er3+ UC emission
has been fitted to a Vial’s type equation, I(t) = A exp( −t
B
)−C exp( −t
D
) [18]. B and D
represent the decay and rise of the transient, respectively; hence, B is essentially the
inverse of 4 S3/2 or 4 F9/2 states lifetime, and D is related to the ETU rate (WETU ) and
Table 5.1: Lifetimes of the 4 F9/2 and 4 S3/2 Er3+ levels in different Y2 O3 : Er3+ , Yb3+
samples detecting at 15129 cm−1 and 17730 cm−1 , respectively, after direct excitation in
the 4 F7/2 level at 20492 cm−1 .
Y2 O3 : Er3+ , Yb3+ Sample τ (4 F9/2 )/µs τ (4 S3/2 )/µs

2%Er3+ , 20%Yb3+ Ball milling 3.9 ± 0.2 1.7 ± 0.1
2%Er3+ , 20%Yb3+ Combustion 2.2 ± 0.1 1.1 ± 0.1
3+ 3+
2%Er , 1%Yb Ball milling 17.6 ± 0.4 4.1 ± 0.2
3+ 3+
2%Er , 1%Yb Combustion 15.3 ± 0.6 2.8 ± 0.2
Figure 5.16: Temporal evolution of the RT 4 F9/2 Er3+ UC luminescence of Y2 O3 : 2%Er3+ , 20%Yb3+
prepared by ball milling (a) and combustion (b) upon excitation at 10256 cm−1 .
1
2
F5/2 , 4 I11/2 lifetimes, according to D = τ (2 F5/2 )
+ τ (4 I111/2 ) + WETU . The best fits for
red and green Er3+ UC luminescence are shown on Tables 5.2 and 5.3, respectively.
The rise of the luminescence intensity after IR excitation is observed in both red
and green UC emissions in all samples. This is a clear proof of an energy transfer
process and evidences the contribution of GSA/ETU mechanism to the total UC
luminescence. In some cases the rise is not starting from zero, this implies that there
is also a contribution of the Er3+ single ion GSA/ESA mechanism; that is, after IR
excitation, an excited Er3+ in the 4 I11/2 or 4 I13/2 level can absorb a second photon
and be promoted to the 4 F9/2 or (2 H11/2 , 4 S3/2 ) levels, respectively. Figure 5.17
shows the temporal evolution of the green Er3+ UC luminescence of Y2 O3 : 2%Er3+ ,
1%Yb3+ prepared by ball milling upon excitation at 10256 cm−1 . Taking into ac-
count that the lifetime of the 4 S3/2 state in the GSA/ESA process should be exactly
the same that after direct Er3+ excitation at 20492 cm−1 , it is possible to estimate
Table 5.2: Lifetimes of the 4 F9/2 Er3+ level in different Y2 O3 : Er3+ , Yb3+ samples mea-
sured detecting at 15129 cm−1 after IR excitation at 10256 cm−1 using a Vial’s type model
(see text).
Y2 O3 : Er3+ , Yb3+ Sample B(4 F9/2 )/µs D(4 F9/2 )/µs

2%Er3+ , 20%Yb3+ Combustion 7.5 ± 0.4 1.7 ± 0.2
3+ 3+
2%Er , 1%Yb Ball milling 163 ± 5 1.7 ± 0.1
2%Er3+ , 1%Yb3+ Combustion 101 ± 7 1.7 ± 0.2
Table 5.3: Lifetimes of the 4 S3/2 Er3+ level in different Y2 O3 : Er3+ , Yb3+ samples mea-
(see text).
Y2 O3 : Er3+ , Yb3+ Sample B(4 S3/2 )/µs D(4 S3/2 )/µs

2%Er3+ , 20%Yb3+ Combustion 1.7 ± 0.2 0.50 ± 0.04
2%Er3+ , 1%Yb3+ Ball milling 122 ± 5 5.0 ± 0.2
2%Er3+ , 1%Yb3+ Combustion 80 ± 10 6.0 ± 0.5
Figure 5.17: Temporal evolution of Er3+ 4 S3/2 UC luminescence intensity in Y2 O3 : 2%Er3+ , 1%Yb3+
obtained by ball milling detecting at 17730 cm−1 upon excitation at 10256 cm−1 . The inset shows a
detail of the intensity rise. The shadow region corresponds to the GSA/ESA contribution taking the
intrinsic lifetime of the 4 S3/2 state after direct excitation.
the GSA/ESA and GSA/ETU contribution to the total intensity. Contributions

of 99% and 1% have been estimated for GSA/ETU and GSA/ESA mechanisms,
respectively [19].
The green (2 H11/2 , 4 S3/2 ) emission lifetimes for Y2 O3 : 2%Er3+ , 1%Yb3+ samples,
observed upon excitation at 10256 cm−1 , are much longer than those obtained for
samples doped with 20%Yb3+ . Even more, lifetimes after IR excitation are longer
than lifetimes after direct Er3+ excitation at 20492 cm−1 . This highlights the energy
transfer in which Yb3+ ions, with a longer lifetime of about 1000 µs, can feed the
4
I11/2 Er3+ multiplet after IR excitation. The same trend is observed for the red
emission decay. The shorter lifetimes detected for samples doped with 20%Yb3+ are
related to a back-transfer from Er3+ to Yb3+ ions whose probability increases with
Yb3+ concentration (see Fig. 5.14).
Conclusions
Y2 O3 : Er3+ , Yb3+ nanocrystalline samples have been synthesized by two different meth-
ods, mechano-chemical synthesis and combustion reaction. A steady-state grain size of 25
nm is reached for the milled Y2 O3 samples while average sizes of 10-15 nm are estimated
for nanoparticles prepared by combustion. However, thermal treatments are required in
order to reduce surface contamination responsible for the luminescence quenching and fi-
nal sizes of ca. 50 nm are obtained for all samples. The nanocrystals have been protected
in capillaries for optical studies and in some cases SiO2 -coated (10 nm thickness). In both
cases, the optical properties are similar indicating the preservation of the nanocrystals
for further surface contamination. The presence of a SiO2 coating surface would allow
to functionalize the nanoparticles for different applications. The dependence of optical
properties on synthesis method or dopant concentration has been studied. The red to
green intensity ratio can be tuned by changing Yb3+ concentration. More intense UC
emission and longer lifetimes have been detected for ball milling samples making them
more suitable for applications. GSA/ETU has been established to be the main mechanism
responsible for the UC luminescence. An enhancement of red emission after IR excitation
has been identified.

Sodium yttrium fluoride, NaYF4 , can crystallize with two different phases, the hexagonal
β-phase (space group P − 6) and the cubic α-phase (space group F m3m). As it will be
discussed later, obtaining NaYF4 : Er3+ , Yb3+ nanoparticles in the pure β phase would
be an important step to get a promising efficient UC material.
Figure 5.18 shows the XRD pattern of NaYF4 nanoparticles co-doped with nominal
concentrations of 2%Er3+ , 20%Yb3+ , synthesized in a planetary ball mill according to the
reaction and the optimal parameters described in Section 3.2, and using different milling
times. It must be noted that when milling time increases, both intensity and width of the
peaks corresponding to the hexagonal phase increase and β-NaYF4 is obtained for times
longer than four hours. On the contrary, XRD peaks matching with cubic phase decrease
when milling time increases. Peaks which do not coincide with any of these phases are
related to the initial YF3 and disappear for times longer than four hours.
Figure 5.18: XRD patterns of NaYF4 : 2%Er3+ , 20%Yb3+ for different milling times (0.5, 2.5, 4.5 and
13 hours). The calculated patterns for hexagonal β-NaYF4 and cubic α-NaYF4 are also shown. Asterisks
indicate YF3 peaks.
The average nanoparticles size is deduced from XRD peaks broadening according to
the Williamson-Hall equation (eq. 4.1). The particle size estimated for NaYF4 : 2%Er3+ ,
20%Yb3+ is about 33 nm after 4 hours milling (Fig. 5.19(a)), and 36 nm after 13 hours
Figure 5.19: Particle size estimation for NaYF4 : 2%Er3+ , 20%Yb3+ nanoparticles prepared in the ball
mill after 4.5 hours (a) and 13 hours (b) milling time. The fitting Williamson-Hall equations are given by
Br cos θ = 0.00467 + 0.00727 sin θ (a) and Br cos θ = 0.00432 + 0.00599 sin θ (b) and the estimated sizes
are L=33 nm (a) and L=36 nm (b), respectively.
Figure 5.20: TEM image of NaYF4 : 2%Er3+ , 20%Yb3+ nanocrystals prepared by ball milling.
(Fig. 5.19(b)). Therefore, it can be seen that from 4 hours, further milling time does
not imply significant changes in particle size. Considering not only the hexagonal phase
formation with the lower cubic contribution, but also the resultant nanocrystals size, it
can be concluded that the optimum milling time for β-NaYF4 : 2%Er3+ , 20%Yb3+ is ca.
10 hours.
A TEM image of NaYF4 : 2%Er3+ , 20%Yb3+ nanocrystals prepared by ball milling

after 13 hours grinding is shown in Fig. 5.20. An average size of ca. 50 nm is estimated,
this value is slightly larger than the one obtained from XRD patterns broadening (36 nm).
Particles are often aggregated forming bigger clusters.
In order to analyze the efficiency of these nanoparticles prepared by ball milling as UC

phosphor material, their optical properties have been compared with the experimental
Figure 5.21: XRD patterns of NaYF4 : 2%Er3+ , 20%Yb3+ bulk and nanoparticles obtained for different
grinding times (1, 2 and 4 hours). The calculated patterns for hexagonal β-NaYF4 and cubic α-NaYF4
are also shown.
results obtained for the bulk β-NaYF4 : 2%Er3+ , 20%Yb3+ synthesized by Krämer [2].
Moreover, NaYF4 : 2%Er3+ , 20%Yb3+ nanoparticles have also been obtained by milling
this bulk β-NaYF4 crystals of 2-5 µm in size in the planetary ball mill. This process
has been carried out using ZrO2 vials and balls, a ball to powder mass ratio of 20:1, an
angular velocity of 300 rpm and stopping 5 minutes every 30 minutes of milling.
Figure 5.21 shows the XRD patterns of the as prepared NaYF4 : 2%Er3+ , 20%Yb3+
nanoparticles for 1, 2 and 4 hour milling and the bulk material. It is observed that the
bulk crystallizes only in the hexagonal system, however, when grinding time increases,
the hexagonal peaks intensity decreases while peaks corresponding to the cubic phase
increase. From hexagonal peaks broadening an average size of 30 nm is obtained for both
1 and 4 hours milling time.
Optical properties
Hexagonal β-NaYF4 doped with Er3+ -Yb3+ and Tm3+ -Yb3+ are up to now, the most
efficient bulk materials for green and blue UC luminescence, respectively [2]. It is an
order of magnitude more efficient than the cubic phase. This can be ascribed to the
low vibrational energies in the hexagonal lattice, phonon cutoff around 350 cm−1 , which
reduces non-radiative multiphonon relaxation processes [20]. Besides, the high emission
efficiency has also been attributed to the structural disorder of the lattice and the presence
of multi-sites for Na+ and RE3+ ions [21], [22]. Both the doping ratio and the phase purity
determine the actual UC efficiency. NaYF4 : 2%Er3+ , 20%Yb3+ colloidal nanocrystals
showed a reduction in the UC efficiency by a factor of 102 -103 [1]. Half of this diminution
was believed to be due to the fact that the nanoparticles crystallize in the α phase, and
the other half to the presence of OH− impurities on the particles surface.
• Synthesis method dependence.
In this work we compare optical properties and UC luminescence efficiency of

NaYF4 : 2%Er3+ , 20%Yb3+ β-bulk and nanoparticles prepared by milling the bulk
hexagonal micropowders, and by the mechano-chemical reaction of the fluorides
described in Section 3.2.
Figure 5.22: RT luminescence spectra of NaYF4 : 2%Er3+ , 20%Yb3+ bulk and nanoparticles (30
nm) upon Er3+ excitation at 26596 cm−1 . Measurements were performed under the same experimental
conditions.
Figures 5.22 and 5.23 show the RT luminescence spectra of bulk β-NaYF4 : 2%Er3+ ,
20%Yb3+ and nanoparticles obtained by milling the bulk powders for 1, 2 and 4
hours (average size of 30 nm) upon direct Er3+ excitation at 26596 cm−1 and af-
ter IR excitation at 10256 cm−1 . Emission bands around 15000 and 18000 cm−1
Figure 5.23: RT UC emission spectra of NaYF4 : 2%Er3+ , 20%Yb3+ bulk and nanoparticles (30
nm) upon IR excitation at 10256 cm−1 . Measurements were carried out under the same experimental
conditions.
detected in both cases are assigned to the red 4 F9/2 → 4 I15/2 and green (2 H11/2 ,
4
S3/2 ) → 4 I15/2 Er3+ transitions, respectively. It is worth mentioning that after vis-
ible excitation into the 4 G11/2 Er3+ level, red emission is more intense than green
luminescence and when milling time increases from 1 to 4 hours, the red to green
intensity ratio decreases by a factor of two. On the other hand, an enhancement of
green emission upon IR excitation is detected for all samples. An evident reduction
in the overall luminescence intensity is observed upon both visible and IR excitation
for longer milling times. The intensity decrease is more noticeable in the UC lumi-
nescence and a reduction in the UC efficiency by a factor of 50 and 103 is estimated
for NaYF4 : 2%Er3+ , 20%Yb3+ nanocrystals obtained after 1 and 4 hours milling,
respectively. The low efficiency observed in nanoparticles is related to two different
factors regarding the milling process. First, the cubic phase emergence which is less
efficient than the hexagonal phase [2]. Second, the particles surface contamination
with the vials and balls which is known to be one of the main problems in mechanical
milling synthesis [23].
Figure 5.24 compares the RT UC luminescence spectra of NaYF4 : 2%Er3+ , 20%Yb3+

obtained using the two described procedures; by milling the hexagonal bulk doped
Figure 5.24: RT UC luminescence spectra of NaYF4 : 2%Er3+ , 20%Yb3+ prepared by milling the
hexagonal bulk (red) and by the mechano-chemical reaction of the fluorides (black) upon IR excitation
at 10256 −1 . Notice the scale factor.
NaYF4 for 1 hour, and by the mechano-chemical reaction of the fluorides (NaF, YF3 ,
YbF3 , ErF3 ) according to equation 3.2 for ten hours. In both cases, nanoparticles
size is around 30-35 nm. It must be pointed out that the sample prepared from the
fluorides reaction is much less efficient (notice the scale factor). The fact that in
this sample the red emission is five times more intense than the green luminescence
is also noteworthy. Since the impurity distribution during the mechano-chemical
reaction is difficult to control, it might be very heterogenous. Therefore, there
is a higher probability of finding nanoparticles with impurity concentration larger
than nominal. This situation makes non-radiative process, responsible for the lu-
minescence quenching, become more important in nanocrystals prepared by the
mechano-chemical reaction.
The RT green (2 H11/2 , 4 S3/2 ) → 4 I15/2 and red 4 F9/2 → 4 I15/2 Er3+ UC luminescence
intensity versus the excitation power density at 10250 cm−1 for NaYF4 : 2%Er3+ ,
20%Yb3+ nanoparticles obtained after 1 hour milling is plotted on a double logarith-
mic scale in Fig. 5.25. A quadratic power dependence is observed below 1 W·cm−2
Figure 5.25: Excitation power dependence of the Er3+ green (a) and red (b) UC emission upon exci-
tation at 10250 cm−1 for NaYF4 : 2%Er3+ , 20%Yb3+ milled for 1 hour.
for both emissions, which is the typical behavior of a two-photon excitation process
in the low-power regime. Hence, this behavior represents the experimental confir-
mation of a two-photon UC process. On the contrary, the slope decreases to 1.0 and
1.2 for green and red luminescence, respectively, at higher excitation power densities
as explained by Pollnau et al. [24] and Suyver et al. [25].
RT temporal evolution of the red and green Er3+ UC luminescence in NaYF4 :

2%Er3+ , 20%Yb3+ bulk and nanoparticles prepared by milling the bulk powders
and by the mechano-chemical reaction of the fluorides have been measured upon
visible and IR excitation (data not shown). The decays measured upon visible exci-
tation at 20492 cm−1 have been fitted to a single exponential. In the experimental
data obtained upon IR excitation at 10246 cm−1 a rise followed by a decay is de-
tected, similarly to the case of Y2 O3 : Er3+ , Yb3+ nanoparticles described previously
in this section. The UC emission decay have been fitted to a Vial’s type equation
[18]. The best fitting values are enclosed on tables 5.4 and 5.5, B and D being the
decay and rise, respectively.
Comparing the temporal evolution of the luminescence intensity for NaYF4 : 2%Er3+ ,
20%Yb3+ bulk and nanoparticles, it can be seen that lifetimes obtained for bulk are
systematically longer than those obtained for nanoparticles, for both red and green
luminescence upon visible and IR excitation. This lifetime decrease associated to
the size diminution is partially related to non-radiative process since, as we have
Table 5.4: Lifetimes of the 4 F9/2 Er3+ level in different NaYF4 : 2%Er3+ , 20%Yb3+ samples
detecting at 15129 cm−1 after visible and IR excitation.
Sample τ / µs B /µs D /µs

(Eexc =20492 cm−1 ) (Eexc =10246 cm−1 ) (Eexc =10246 cm−1 )
Bulk 385 ± 5 655 ± 5 60 ± 2
1 hour milled 145 ± 5 580 ± 10 230 ± 5
9.5 hours milled 54 ± 4 78 ± 5 5.8 ± 0.5
Table 5.5: Lifetimes of the 4 S3/2 Er3+ level in different NaYF4 : 2%Er3+ , 20%Yb3+ samples

(Eexc =20492 cm ) (Eexc =10246 cm ) (Eexc =10246 cm−1 )
−1 −1
Bulk 116 ± 2 316 ± 5 42 ± 2

1 hour milled 33 ± 1 177 ± 6 60 ± 3
9.5 hours milled 11.5 ± 0.5 23.5 ± 0.8 2.3 ± 0.2
seen, it is accompanied by an intensity decrease. However, the intensity decrease

is notably more important than the lifetime reduction. As an example, in NaYF4 :
2%Er3+ , 20%Yb3+ milled for 1 hour the green UC luminescence intensity diminishes
by a factor of 50 while the lifetime is only reduced to half its value. This additional
intensity reduction should be related to the cubic phase appearance.
Conclusions
NaYF4 : 2%Er3+ , 20%Yb3+ nanocrystals have been prepared by two different methods,
and particles of around 30 nm in size are obtained in all cases. The sample obtained
after milling the bulk for 1 hour shows the best UC luminescence efficiency but it is still
much less efficient than bulk. This is related to the emergence of cubic α-NaYF4 , and to
surface contamination of the nanocrystals during the milling process, associated to the
larger surface-to-volume ratio in nanoparticles.
5.2.3 Tb3+ or Eu3+ and Yb3+ co-doped Gd3 Ga5 O12 and Y3 Al5 O12
Gadolinium gallium garnet, GGG, and yttrium aluminum garnet, YAG, are suitable mate-
rials as host for luminescent trivalent lanthanide ions. They belong to the cubic crystalline
system, having a garnet structure, with Ia3d space group (Fig. 5.26). The bulk cell pa-
rameters are a0 =12.376 Å and a0 =12.010 Å for GGG and YAG, respectively. For both
systems, the garnet structure is composed of a 24(c) dodecahedral site (D2 point symme-
try) for Gd3+ or Y3+ depending on the case with a coordination number 8, and two sites
for Ga3+ and Al3+ , respectively, a 16(a) octahedral site with a coordination number 6,
and a 24(d) tetrahedral site of coordination number 4. O2− ions occupy the 96(h) sites
with each one being a member of two dodecahedra, one octahedron and one tetrahedron.
The garnet structure can be viewed as interconnected dodecahedra, octahedra and tetra-
hedra with shared O atoms at the corners of the polyhedra. Taking into account the ionic
radii, the doping RE3+ impurities are expected to enter into the Gd3+ or Y3+ sites which
possess D2 symmetry [26].
Figure 5.26: GGG (left) and YAG (right) cubic structures. In GGG structure the two Ga3+ sites
with octahedral and tetrahedral coordination is shown, while the Y3+ site with coordination number 8
is depicted in YAG structure.
GGG and YAG nanocrystalline powders co-doped with nominal concentrations of

2%Tb3+ -5%Yb3+ , 2%Eu3+ -5%Yb3+ , and 2%Eu3+ -1%Er3+ have been prepared by the sol-
gel Pechini’s method described in Section 3.4. It is worth mentioning that lanthanides
replace Gd3+ or Y3+ ions, hence, the amount of RE3+ doping ions (1%, 2% or 5% mol)
is with respect to Gd3+ or Y3+ . All the prepared samples have been analyzed using the
XRD technique.
Figure 5.27 shows the XRD pattern of GGG: 2%Eu3+ , 5%Yb3+ . The Rietveld re-
finement of the GGG nanopowders XRD pattern is consistent with a cubic garnet single
phase (space group Ia3d, a=12.391 Å) but contains small traces (up to 5%) of a second
phase, Gd3 GaO6 . The obtained lattice constant is similar to that found for an Eu3+
doped GGG nanocrystalline sample prepared by the same Pechini procedure [27]. From
Rietveld refinement and considering XRD peaks broadening, the average size of the crys-
tallite grains has been determined to be ca. 30 nm. The Eu3+ -Er3+ doped GGG samples
are also cubic garnet single phase apart from a small trace contamination of Gd3 GaO6
phase, as found for the Eu3+ -Yb3+ doped samples.
Figure 5.27: XRD pattern of GGG: 2%Eu3+ , 5%Yb3+ prepared following Pechini’s method (a) and
Rietveld fit (b). The asterisk indicates the main peak of the Gd3 GaO6 impurity phase.
Figure 5.28 presents the XRD pattern of YAG: 2%Eu3+ , 5%Yb3+ together with the
Rietveld fit. Both the Eu3+ -Yb3+ and Eu3+ -Er3+ doped nanocrystalline YAG samples
crystallize only in the cubic garnet phase (space group Ia3d, a=12.031 Å) without evi-
dence of contamination from other phases. The corresponding size of the nanoparticles
obtained from Rietveld fitting is around 40 nm. There is no evidence of extra peaks due
to phase segregation of the doping components.
TEM images of GGG and YAG nanocrystalline samples (see Fig. 5.29) show that the
samples are made of particles of different shapes and sizes. Most of them are spherical
with sizes in the 30-50 nm range but in some cases particles are aggregated forming larger
clusters.
Figure 5.28: XRD pattern of YAG: 2%Eu3+ , 5%Yb3+ prepared by the sol-gel Pechini’s method (a)
and Rietveld fit (b).
Figure 5.29: TEM images of GGG: 2%Tb3+ , 5%Yb3+ prepared by the sol-gel Pechini’s method. In
the HRTEM image, the observed interplanar distances (0.30 nm) corresponds to the (012) crystal plane.
Figure 5.30 shows IR absorption spectra of both host materials, GGG and YAG, be-
tween 300 and 1300 cm−1 . Analyzing this frequency range, information about vibrational
energy of each matrix can be obtained. The higher energy absorption band is observed
at around 680 cm−1 for the GGG, while it is shifted to higher energies and appears at
around 800 cm−1 in YAG. The highest energy phonons belong to the Ga-O and Gd-O
Figure 5.30: IR absorption spectra of GGG: 2%Eu3+ , 5%Yb3+ and YAG: 2%Eu3+ , 5%Yb3+ synthesized
by Pechini’s method. Nanopowders were diluted in KBr to perform these measurements.
or Y-O and Al-O stretching vibration modes for GGG and YAG, respectively [28]. Ac-
cording to the gap’s law, multiphonon relaxation becomes much more probable for higher
phonon energies. Therefore, non-radiative processes, responsible for the diminishing of
luminescence, are more likely to happen in YAG than in GGG.
Optical properties
GGG and YAG are host materials with interesting optical properties [29]. Average phonon
energy in sesquioxides is quite low for oxide materials, which is especially beneficial for
optical applications. YAG: Nd3+ is one of the most important materials used for solid-
state lasers. GGG has also shown to have several advantages over other laser materials
[30]. A detailed investigation on the spectroscopy and excited state dynamics at RT has
been carried out on Tb3+ -Yb3+ and Eu3+ -Yb3+ co-doped GGG and YAG.
• Tb3+ -Yb3+ System.
Figure 5.31 shows Tb3+ excitation spectra in both GGG and YAG co-doped with
2%Tb3+ , 5%Yb3+ nanoparticles. In addition to the f−f Tb3+ transitions from the
ground state 7 F6 to different excited states plotted in the Dieke diagram, the excita-
tion spectrum of YAG: 2%Tb3+ , 5%Yb3+ shows two intense bands at around 31060
and 36100 cm−1 . We assign these transitions to the spin-allowed and spin-forbidden
4f−5dt transitions of Tb3+ , in agreement with the data reported by Dorenbos for
Figure 5.31: RT excitation spectra of GGG: 2%Tb3+ , 5%Yb3+ and YAG: 2%Tb3+ , 5%Yb3+ recording
Tb3+ luminescence at 18420 cm−1 .
the same compound (36500 and 30860 cm−1 ) [31], [32]. The high energy band is
more than one order of magnitude stronger than the low energy band. Surprisingly,
these bands are not detected in the case of GGG: 2%Tb3+ , 5%Yb3+ pointing out a
different crystal-field strength on the Tb3+ site.
Figure 5.32 presents Tb3+ 5 D4 → 7 FJ emission upon UV excitation in the 5dt state
at 37040 cm−1 for GGG: 2%Tb3+ , 5%Yb3+ and YAG: 2%Tb3+ , 5%Yb3+ nanocrys-
talline samples. As expected from excitation spectra, Tb3+ luminescence is one
order of magnitude more intense in YAG than in GGG.
Figure 5.33 compares the RT luminescence spectra of nanocrystalline GGG: 2%Tb3+ ,

5%Yb3+ and YAG: 2%Tb3+ , 5%Yb3+ upon visible and IR excitation. The first is
obtained exciting directly the Tb3+ ions at 37040 cm−1 while the UC luminescence
is observed after the excitation of Yb3+ ions at 10250 cm−1 . In all cases, the emis-
sion bands observed in the red-blue region (15000-21000 cm−1 ) are assigned to the
transitions from the 5 D4 multiplet to lower energy 7 FJ states of Tb3+ ions (see Fig.
5.34). Yb3+ pairs luminescence is not detected in our samples.
The presence of the parity allowed 4f−5dt transitions in the near-UV region in the
case of YAG: 2%Tb3+ , 5%Yb3+ , and additionally, the demonstration of Tb3+ -Yb3+
UC luminescence, opens the possibility to use this system also as a downconversion
Figure 5.32: RT luminescence spectra of GGG: 2%Tb3+ , 5%Yb3+ and YAG: 2%Tb3+ , 5%Yb3+ upon
excitation at 37040 cm−1 . Measurements were carried out under the same experimental conditions.
Figure 5.33: RT luminescence spectra of GGG: 2%Tb3+ , 5%Yb3+ exciting at 37040 cm−1 (a) and
10250 cm−1 (b). RT emission spectra of YAG: 2%Tb3+ , 5%Yb3+ exciting at 37040 cm−1 (c) and 10250
cm−1 (d). * is an artifact and represents the laser at twice the excitation frequency.
Figure 5.34: Energy level scheme of Tb3+ and Yb3+ ions with the 5 D4 →7 FJ and the 5 D3 →7 FJ
Tb3+ luminescent transitions and the cooperative sensitization mechanism.
phosphor. In this process, one Tb3+ ion would generate two Yb3+ ions in the excited
state [33]. This could be applied for the improvement of solar cells efficiency. This is
particularly important taking into account that the 4f−5dt bands in YAG: 2%Tb3+ ,
5%Yb3+ are located close to band gap of clear glasses usually integrated in front of
the solar cell panels as protectors.
As we have seen, the excitation spectrum of Tb3+ in YAG (Fig. 5.35(a)) consists
of two strong bands, at 36100 cm−1 and 31060 cm−1 , compared to the intensity
of the f−f transitions. Excitation on the 4f−5dt bands is followed by a fast non-
radiative relaxation to the 5 D3 and 5 D4 excited states, from where emission takes
place (Fig. 5.35(b)). Although green 5 D4 → 7 FJ Tb3+ UC luminescence has been
widely investigated, there are only few examples of blue Tb3+ UC emission from
the 5 D3 multiplet in the literature [34], [35]. Luminescence from this state between
22000 and 27000 cm−1 has been observed after IR excitation of YAG: 2%Tb3+ ,
5%Yb3+ (Fig. 5.35(c)).
The only way to access the 5 D3 excited state is via a three photon process (see
Figure 5.35: RT excitation spectrum of YAG: 2%Tb3+ , 5%Yb3+ (a), RT luminescence spectrum after
excitation at 37037 cm−1 (b), and RT emission spectrum after excitation at 10250 cm−1 (c). Note the
scale factors.
Fig. 5.34). In order to prove this assumption, the power dependence of the emitted
photons of the 5 D4 → 7 FJ and 5 D3 → 7 FJ Tb3+ transitions upon IR excitation has
been measured (Fig. 5.36). Slopes of 2.6 and 3.3 are obtained for 5 D4 → 7 FJ and 5 D3
→ 7 FJ Tb3+ emissions, respectively. With this result, we can propose that a three
photon process is also involved in the green 5 D4 → 7 FJ luminescence, indicating
the relevance of a non-radiative relaxation from the 5 D3 to 5 D4 states. Such a three
photon process for the green UC Tb3+ emission was previously observed [36]. This
is in agreement with a possible CR process populating the 5 D4 from the 5 D3 state
[37].
• Eu3+ -Yb3+ System.
Figure 5.37 presents Eu3+ luminescence upon excitation at 25450 cm−1 for GGG:
2%Eu3+ , 5%Yb3+ and YAG: 2%Eu3+ , 5%Yb3+ nanoparticles under the same exper-
imental conditions. It can be observed that emission intensity is of the same order of
magnitude for both cases. Emission features are different for Eu3+ -Yb3+ co-doped
5
GGG and YAG nanoparticles. D0 → 7 F6,5 Eu3+ emission peaks are shifted to
Figure 5.36: Excitation power dependence of the Tb3+ 5 D4 → 7 FJ luminescence (a) and Tb3+ 5 D3 →
7
FJ emission (b) in YAG: 2%Tb3+ , 5%Yb3+ upon 10250 cm−1 excitation.
higher energies for GGG, 5 D0 → 7 F3,2 transitions have very diverse shape in GGG
and YAG, and 5 D0 → 7 F1 transitions show different energy separation between lines.
Figure 5.37: RT emission spectra of GGG: 2%Eu3+ , 5%Yb3+ and YAG: 2%Eu3+ , 5%Yb3+ upon
Figure 5.38 shows the Stokes and anti-Stokes RT luminescence spectra of nanocrys-
talline GGG: 2%Eu3+ , 5%Yb3+ and YAG: 2%Eu3+ , 5%Yb3+ . Eu3+ emission is
obtained upon direct excitation at 25450 cm−1 , whereas Eu3+ UC luminescence
is detected upon Yb3+ ions excitation at 10250 cm−1 . Unlike Yb3+ to Tb3+ UC
luminescence, there are only few examples of Yb3+ to Eu3+ UC emission in the
literature [13], [14], [38]. The bands observed in the spectra are assigned to 5 D0 →
7
FJ Eu3+ transitions (see Fig. 5.39). In the UC luminescence spectra (Fig. 5.38(b)
and 5.38(d)), emission bands in the 14500-15500 cm−1 range due to Er3+ impurities
are detected in addition to the Eu3+ f−f transitions. This fact was also observed by
Tanner [38] even with starting materials with purity of 99.999%. In order to under-
stand if Er3+ impurities play any role in the Yb3+ to Eu3+ UC luminescence process,
GGG: 2%Eu3+ , 1%Er3+ and YAG: 2%Eu3+ , 1%Er3+ nanocrystalline samples have
also been investigated.
Figure 5.38: RT luminescence spectra of GGG: 2%Eu3+ , 5%Yb3+ exciting at 25450 cm−1 (a) and
10250 cm−1 (b). RT emission spectra of YAG: 2%Eu3+ , 5%Yb3+ exciting at 25450 cm−1 (c) and 10250
cm−1 (d). The emission corresponding to Er3+ transitions are shown between arrows.
Figure 5.40 compares the UC luminescence of GGG: 2%Eu3+ , 1%Er3+ upon excita-
tion of Er3+ ions with a LD at 10250 cm−1 with that of GGG: 2%Eu3+ , 5%Yb3+ .
Er3+ excitation at 10250 cm−1 (Fig. 5.40 (a)) is able to induce UC luminescence
bands observed at around 15000 and 18000 cm−1 , assigned to the 4 F9/2 → 4 I15/2 ,
and (2 H11/2 , 4 S3/2 ) → 4 I15/2 Er3+ transitions, respectively, with the absence of Eu3+
emission. It is concluded that excitation at 10250 cm−1 induces Er3+ luminescence
but it is not able to induce Eu3+ luminescence.
Figure 5.39: Energy level diagram of Eu3+ and Yb3+ ions with the 5 D0 →7 FJ Eu3+ emissions and
the proposed Yb3+ -Eu3+ energy transfer mechanisms.
Figure 5.40: RT UC luminescence spectra of GGG: 2%Eu3+ , 1%Er3+ (a) and GGG: 2%Eu3+ , 5%Yb3+
(b) exciting at 10250 cm−1 .
Therefore, any relevant role of Er3+ in the Yb3+ to Eu3+ UC luminescence is ex-
cluded, although the efficient Yb3+ to Er3+ energy transfer compared to Yb3+ to
Eu3+ one is pointed out.
The time evolution of the 5 D4 →7 FJ Tb3+ and 5 D0 →7 FJ Eu3+ emission in GGG

and YAG singly-doped and co-doped nanoparticles has been recorded after direct
excitation at around 20365 and 16920 cm−1 for Tb3+ and Eu3+ , respectively. GGG
experimental decays are shown in Fig. 5.41.
Figure 5.41: Temporal evolution of 5 D4 → 7 FJ Tb3+ RT luminescence in GGG: 2%Tb3+ , 5%Yb3+

upon excitation at 20365 cm−1 (a), and 5 D0 → 7 FJ Eu3+ RT emission in GGG: 2%Eu3+ , 5%Yb3+ upon
excitation at 16920 cm−1 (b).
Table 5.6: Lifetimes of the 5 D4 Tb3+ and 5 D0 Eu3+ levels in GGG and YAG samples
measured at the emission energy, Eem , after direct excitation into the emitting levels.
Eem /cm−1 Sample τ /ms Sample τ /ms

18396 GGG: 2%Tb3+ , 5%Yb3+ 4.2 ± 0.2 GGG: 2%Tb3+ 4.1 ± 0.2
18396 YAG: 2%Tb3+ , 5%Yb3+ 4.9 ± 0.2 YAG: 2%Tb3+ 4.7 ± 0.2
14134 GGG: 2%Eu3+ , 5%Yb3+ 4.0 ± 0.2 GGG: 2%Eu3+ 4.1 ± 0.2
14090 YAG: 2%Eu3+ , 5%Yb3+ 4.6 ± 0.2 YAG: 2%Eu3+ 4.6 ± 0.2
Experimental data have been fitted to a single exponential for all samples and the
corresponding lifetimes are shown in Table 5.6.
The Tb3+ and Eu3+ lifetimes obtained in this work are longer than the average
values measured in doped Y3 GaO6 [39] or Li2 SiO3 [40] nanoparticles. It is also
worth mentioning that the lifetimes for Tb3+ and Eu3+ ions in nanocrystalline GGG
and YAG are significantly longer than the usual values found for the doped bulk
oxide counterparts [41]. This behavior was also observed for nanocrystalline Eu3+
doped Y2 O3 [42] and ZrO2 [43] powders and it is compatible with the small particle
size of the nanopowders, around 40 nm. In fact, this lengthening of the lifetimes
can be ascribed to a lower refractive index (neff ) surrounding the lanthanide ions
in the nanocrystalline material, compared to the micrometer size host, according
to a higher surface-to-volume ratio [42]. The lifetimes obtained by fitting the 2 F5/2
→ 2 F7/2 Yb3+ luminescence decay to a single exponential in co-doped GGG and
Table 5.7: Lifetimes of the 2 F5/2 Yb3+ excited state in GGG and YAG samples measured
detecting at 9750 cm−1 upon excitation at 10250 cm−1 .
Sample τ /µs
GGG: 2%Tb3+ , 5%Yb3+ 595 ± 10
YAG: 2%Tb3+ , 5%Yb3+ 485 ± 10
GGG: 2%Eu3+ , 5%Yb3+ 555 ± 10
YAG: 2%Eu3+ , 5%Yb3+ 480 ± 10
YAG samples are shown in Table 5.7. The same lifetime values are observed for
singly Tb3+ (or Eu3+ ) doped samples and Tb3+ -Yb3+ (or Eu3+ -Yb3+ ) co-doped
ones. Hence, there is no evidence of back-transfer from Tb3+ or Eu3+ to Yb3+ .
Tb3+ and Eu3+ UC luminescence has been measured in GGG and YAG upon exci-
tation at 10250 cm−1 in the 2 F5/2 Yb3+ level. Tb3+ and Eu3+ ions have no excited
states resonant with the 2 F7/2 → 2 F5/2 Yb3+ transition in the IR region. Therefore,
the UC emission cannot be explained by a direct transfer (GSA/ETU) from the
excited Yb3+ to a Tb3+ or Eu3+ ion. A cooperative GSA/ESA mechanism in Tb3+ -
Yb3+ dimers coupled by exchange was demonstrated to be responsible for Tb3+ UC
emission after Yb3+ excitation in Cs3 Tb2 Br9 : Yb3+ at low temperatures [12]. How-
ever, GGG and YAG crystal structures are not able to accommodate Tb3+ -Yb3+ or
Eu3+ -Yb3+ dimers. Hence, GSA/ESA in dimers can be ruled out as a mechanism
responsible for the UC luminescence. Additionally, UC phenomena in YF3 : Tb3+ ,
Yb3+ were explained as a cooperative sensitization [10]. Strek et al. proposed two
mechanisms to explain the anti-Stokes Eu3+ emission in KEu0.2 Yb0.8 (WO4 )2 upon
Yb3+ excitation, which were cooperative sensitization and a process based on Yb3+
pairs [13].
The unambiguous manner to investigate the mechanism involved in the UC lu-

minescence is to study its temporal evolution under pulsed excitation [12]. The
cooperative sensitization, a three ions process, involves energy transfer from two
sensitizers (Yb3+ ions) to the activator (Tb3+ or Eu3+ ). This is a slow process
which is active during the Yb3+ lifetime. The temporal evolution of the 5 D4 → 7 FJ
Figure 5.42: Temporal evolution of Tb3+ UC emission intensity in YAG: 2%Tb3+ , 5%Yb3+ detecting
at 18396 cm−1 after pulsed excitation at 10250 cm−1 .
green UC luminescence of Tb3+ obtained upon Yb3+ excitation at 10250 cm−1 for
the YAG: 2%Tb3+ , 5%Yb3+ sample is shown in Fig. 5.42. The rise of the emis-
sion intensity after IR excitation is a clear fingerprint of an energy transfer process.
The cooperative sensitization is proposed as the mechanism responsible for Tb3+
UC luminescence in GGG and YAG systems co-doped with Yb3+ . This mecha-
nism involves three ions and is depicted in Fig. 5.34. The cooperative sensitization
responsible for the Tb3+ UC luminescence is as follows: two Yb3+ ions in the ex-
cited state transfer non-radiatively the energy to the 5 D4 states of Tb3+ . The same
cooperative sensitization mechanism would be expected in the Eu3+ -Yb3+ UC lu-
minescence (see Fig. 5.39). The Tb3+ lifetime observed in the UC decay transients,
1.8 ± 0.1 ms, is shorter than the one obtained after direct excitation, 4.9 ± 0.2 ms.
(See Fig. 5.42 and Table 5.6). This difference is related to the fact that not all Tb3+
have necessarily an Yb3+ ion as a close neighbor. In UC measurements, only those
Tb3+ which are close to Yb3+ ions are excited, whereas after direct excitation, all
Tb3+ are involved in the luminescence [44].
The temporal evolution in Tb3+ -Yb3+ systems can be simulated by considering a

cooperative sensitization mechanism. Taking into account the four level system
schematically shown in Fig. 5.43, we can write down the four coupled differential
rate equations (eq. 5.1) describing the populations of each level, Ni :
Figure 5.43: Diagram of the cooperative sensitization of Tb3+ UC luminescence. In this process two
excited Yb3+ ions transfer non-radiatively their energy to Tb3+ ions.
dN0 N1
= −GN0 + + 2kCS N12 N2
dt τ1
dN1 N1
= GN0 − − 2kCS N12 N2
dt τ1
dN2 N3
= − kCS N12 N2 (5.1)
dt τ3
dN3 N3
= − + kCS N12 N2
dt τ3
where τ1 and τ3 represent the Yb3+ 2 F5/2 and Tb3+ 5 D4 experimental lifetimes, re-
spectively, G is the power dependence GSA rate constant and kCS is the cooperative
sensitization rate parameter. The physical meaning of these equations is the follow-
ing; the rise is related with half of the Yb3+ 2 F5/2 lifetime, that would correspond to
the lifetime of Yb3+ pairs, whereas the decay is related to the Tb3+ 5 D4 lifetime. N0
and N2 are known from Yb3+ and Tb3+ concentration values, respectively, whereas
N1 and N3 are an initial hypothesis since they cannot be experimentally measured
in an accurate manner. Values of 485 ± 10 µs (see Table 5.7) and 1.8 ± 0.1 ms
(see Fig. 5.42) have been obtained from independent measurements for τ1 and τ3 ,
respectively, in YAG: 2%Tb3+ , 5%Yb3+ . It is possible to obtain the only unknown
parameter, kCS , by fitting the time dependent evolution of the UC luminescence,
obtained experimentally after short-pulsed excitation into the Yb3+ 2 F5/2 level, to
this set of equations (Fig. 5.42). In this way, a value of 1350 s−1 has been determined
for the energy transfer rate constant, kCS N1 N2 .
UNIVERSIDAD DE CANTABRIA
Facultad de Ciencias
Departamento de Fı́sica Aplicada
Tesis Doctoral
SYNTHESIS, STRUCTURAL CHARACTERIZATION AND

SPECTROSCOPIC STUDY OF NANOCRYSTALLINE
AND MICROCRYSTALLINE MATERIALS
Rosa Martı́n Rodrı́guez

Chapter 2
Theory
2.1 Introduction
The models and approximations utilized for the interpretation of the experimental data are
presented along this chapter. Firstly, the increasing interest of nanomaterials regarding
both fundamental studies and applications is outlined. Secondly, the determination of the
energy levels of an ion in a crystal, and the transitions among these levels, responsible for
the RE and TM ions optical properties, are described. From Fermi’s Golden Rule which
defines the transition probability, concepts like the transition lifetime are introduced.
Non-radiative relaxation processes, responsible for most of the luminescence quenching,
and the different UC mechanisms, as well as the basic concepts of Raman spectroscopy are
also explained. Moreover, the size dependence of the optical properties in semiconductor
nanoparticles is deduced. Finally, structural phase transitions are briefly described.
2.2 Interest of nanoparticles
Remarkable examples of nanotechnology throughout history are the famous Lycurgus Cup
which shows dichroism due to the interaction of light with metal nanoparticles embedded
in a glass matrix, or the addition of metallic constituents to glass, in order to create
stained glass windows [1]. Nowadays, nanoscience and nanotechnology deal with the study
of materials with size in the nanometer range, which is extremely interesting for several
reasons. Firstly, relevant physical laws change at the nanometric scale, this is, many of the
classical laws are not valid anymore and quantum mechanics must be applied. Besides
9
10 2.2. Interest of nanoparticles
quantum confinement effects [2], changes with size reduction are related to the higher
surface-to-volume ratio in nanoparticles [3], [4]. Secondly, some metastable phases are only
observed in reduced-size materials [5]. Thirdly, diverse areas of technology have pointed
their attention in the last years to make smaller components not only for electronics but
also in other fields like new materials, health-sciences or renewable energy sources [1], [6].
The well-known Moore’s law summarizes the scale to get the same function by making the
working elements smaller. The tunnel diode is the first example in which the appearance
of quantum physics at small sizes led to a new device. The fact that the energy gap of QDs,
and therefore the emitting wavelength, is blue-shifted when particle size decreases, is of
great interest for solar cells and bio-labeling [7], [8]. Carbon black which contains carbon
nanotubes, has been used for a long time as an additive to the rubber in automobile tires.
Other examples in which the industry has taken advantage of nanomaterials are titanium
dioxide pigment in white paint, large synthetic biomolecules or nanoparticle slurries for
polishing.
Moving to our research field, there is a considerable interest in the development of

highly-luminescent nanomaterials for biological applications such as bio-labeling, drug
delivery or diagnostics. Metallic and magnetic nanoparticles with the adequate organic
capping, and core-shell semiconductor nanocrystals, have been widely used in biological
experiments [8], [9], [10]. Water-soluble nanocrystals as well as materials with appropri-
ate optical properties are needed. In this sense, the study of the optical properties of
RE-doped nanoparticles for biological applications is increasing because of lanthanides
properties such as narrow absorption and emission bands, high quantum efficiency, and
long lifetimes [11], [12], [13]. Different nanomaterials find also applications in renewable
energy sources such as photovoltaic solar cells technology [7], [14]. UC and downconversion
processes in nanoparticles doped with RE or TM ions may also be implemented for this
purpose. For instance, NaYF4 : Er3+ phosphors have been proposed for the enhancement
of solar cells response [15], [16].
2. Theory 11
2.3 Insulating materials. Optical properties

2.3.1 Doped insulators
Insulators tend to be colorless since they are characterized by a transparency range in the
ultraviolet (UV)- visible - near infrared (NIR) range. The sharp drop in the transmission
for λ > 6000 nm is caused by vibrational absorption, while in the UV spectral region, λ <
200 nm, it is due to absorption in the material bandgap. Insulating hosts can present new
optical properties thanks to the presence of dopant impurity ions, such as RE or TM ions
[17]. The properties of the optically active ions may be understood according to the idea
of a coordination complex, ABn . It is a structure consisting of a central cation A bonded
to the n closest neighbors B, called ligands.
2.3.2 Ions in a static crystalline environment
According to a point-charge model, the valence electrons belong to the dopant ion A,
and the effect of the lattice modifies its energy levels by the influence of the ligand ions
B through the electric field that they produce at the site of A. This electrostatic field is
called the crystal-field, and the Hamiltonian can be written as [18], [19], [20]:
H = HF I + HCF (2.1)
HF I being the free ion A Hamiltonian, and HCF the crystal-field Hamiltonian:
HF I = H0 + H 0 + HSO (2.2)
where H0 includes the kinetic and potential energy in the central-field approximation,
due to the nucleus and the inner closed-shell electrons, H 0 corresponds to the Coulomb
interaction among the outer electrons, and HSO describes the interaction between the
magnetic moment of the electron, s, and its angular moment, l, and it is known as the
spin-orbit coupling.
Depending on the relative importance of the different terms, diverse approaches can
be considered in order to diagonalize the Hamiltonian:
• Weak crystal field : HCF HSO , H 0 , H0 . In this regime the energy levels of the free
ion A are only slightly perturbed by the crystal field. HCF is taken as a perturbation
over the 2S+1 LJ states (J = L + S). It is appropriate to describe the energy levels of
trivalent RE ions, since for these ions the optically active 4f electrons are screened
by the outer filled 5s2 5p6 sub-shells. These electrons partially shield the crystal field
created by the B ions.
• Intermediate crystal field : HSO HCF < H 0 . In this case the spin-orbit interaction
which is weaker than the crystal field, is initially neglected. The crystal field is
2S+1
considered a perturbation on the L terms.
• Strong crystal field : HSO < H 0 < HCF . In this approach the crystal field term
dominates over the other two. The electron-electron and the spin-orbit interaction
are taken into account as a perturbation. This applies to 3d TM ions and 4f-5d
inter-configurational transitions in RE ions.
2.3.3 The configurational coordinate diagram
So far, we have considered that the optical center A is embedded in a static lattice.
However, in a real crystal, the environment of A is not static but dynamic. In order
to understand the effect of the crystalline vibrations on the optical spectra, we must
consider that the optically active ion A is coupled to the vibrating lattice [18]. The
configurational coordinate model (Fig. 2.1) represents the energy curves in the ground
and excited electronic states as a function of an effective normal mode coordinate. In the
simplest case, this model which can account for such a type coupling, is based on several
approximations; the Franck-Condon approximation, the harmonic approximation and the
fact that only one vibration mode is considered [20].
A = −KQ0 is the electron-lattice coupling constant, where K is the force constant,
considered equal for the ground and excited states for simplicity, and Q0 is the minimum
of the excited state parabola. The sign of A determine whether the equilibrium distance
is longer in the excited state or in the ground state, and it can be obtained from the
pressure dependence of the absorption or emission spectra [21]. The shift between the
2. Theory 13
Figure 2.1: Single configurational coordinate diagram for the AB6 center.
ground and excited states configurational curves is larger for bigger values of A. This
shift which causes a displacement between emission and absorption maxima, is usually
quantified by the Huang-Rhys parameter, S, defined as:
A2 KQ20
S= = (2.3)
2K~ω 2~ω
S is a measure of the linear coupling between the electronic excited state and the
vibrational mode, and it is related with the different equilibrium geometry between ground
and excited states, with S 1 for weak coupling, S between 1 and 4 for medium coupling
and S > 4 for strong coupling. The energy difference between the maximum of the
absorption and emission bands, ESS , is called Stokes shift:
ESS = 2S~ω (2.4)
Then, from the experimental Stokes shift, the Huang-Rhys parameter, and therefore
the distortion of the excited state with respect to the fundamental one, can be obtained.
The stronger the linear coupling is, the larger the bandwith, the Huang-Rhys parameter
and the Stokes shift are. f-f transitions in RE3+ ions show weak coupling with the lattice
due to the mentioned shield of the 4f electrons, hence, the configurational energy curves
are one above the other, and optical spectra consist of fine lines due to pure electronic
transitions with small Stokes shift. On the contrary, d-d transitions in TM ions may also
show strong coupling giving rise to broad bands in their spectra.
2.3.4 Light absorption and emission processes
Determining the energy levels of an optically active center is important because optical
spectra result from transitions among these energy levels. Different processes can take
place such as absorption, due to transitions from lower energy levels to higher energy
levels, and emission, in which the final state is lower in energy than the initial one.
Considering a two-level system consisting of an initial state |i > and a final state |f >
separated by an energy Ef − Ei = ~ωf i , the probability of inducing an optical transition
from the state |i > to the state |f > upon the absorption of a photon is given by the
Fermi’s Golden Rule [20]:
2π
Wi→f = | < i|Hint |f > |2 ρk (ωf i ) (2.5)
~2
where Hint is the radiation-matter interaction Hamiltonian and ρk (ωf i ) is the density
of radiation field modes per unit angular velocity. The parity and spin selection rules
governing a transition are determined by the matrix element < i|Hint |f >.
Relaxation from excited states can take place by both radiative and non-radiative
processes. Purely radiative |f >→ |i > stimulated emission presents the same transition
probability as the inverse absorption process, Wf →i = Wi→f [22]. When spontaneous
emission is the radiative desexcitation channel between |f > and |i > states, the transition
probability is given by the Einstein coefficient, Wfsp→i = Af i . The rate at which the
population of the |f > state, Nf , decays by spontaneous emission in the absence of non-
radiative processes is given by:
dNf
= −Nf Af i (2.6)
dt
t
Nf (t) = Nf (0) exp −
τrad
1
where τrad is the radiative lifetime of the spontaneous emission, τrad = Af i
.
2. Theory 15
Figure 2.2: Classical interpretation of the activation energy, Ea , in non-radiative transitions.
2.3.5 Non-radiative transitions
Contrary to RE ions, in which luminescence is a quite common phenomenon even from

multiple excited states, non-radiative relaxation processes usually dominate the radiative
ones in TM systems. A distinction is usually made between non-radiative processes within
a given complex; multiphonon relaxation, and those involving more than one optically
active center; energy transfer processes [18], [22].
• Multiphonon relaxation: In these processes, the electronic excitation is totally or

partially transformed into vibrational energy within a fempto-picoseconds timescale.
Multiphonon relaxation usually occurs down to the lowest energy excited state,
therefore, if luminescence exists, it takes place from this state. This is known as
Kasha’s rule, and RE compounds are exceptions to this rule. The configurational
coordinate model (Fig. 2.2) provides a qualitative explanation to describe the tem-
perature dependence of multiphonon relaxation from an excited state |f > to some
lower state |i > in TM systems. According to a model proposed by Mott, the non-
radiative desexcitation probability is associated to an activation energy, Ea , and is
given by [23]:

Ea
Wnr f →i = A exp − (2.7)
kT
where Ea is the energy difference between the minimum of the excited state Ef (Q)
parabola and the Ei (Q) - Ef (Q) parabolas crossing point (Fig. 2.2). However, mul-
tiphonon relaxation can also be studied quantically considering tunneling processes
between |f > and |i > states.
In the weak coupling case which is characteristic for RE3+ compounds, the transition
rate between |f > and |i > states through multiphonon relaxation can be expressed
as:
1
Wnr f →i (T = 0K) ∝ exp(−βp) = (2.8)
τnr
This is the so-called energy gap’s law, where p is the reduced energy gap and rep-
resents the number of phonons involved in the relaxation process [24]. For f → f
transitions in RE ions, multiphonon relaxation processes are dominant for energy
separation between levels involving p ≤ 6.
• Energy transfer : All or part of the energy of an initially excited center is transferred
to another center and then either emitted or converted into heat by multiphonon
relaxation. In order to simplify the analysis of this kind of mechanisms, a single
energy transfer step between two ions is considered:
S ∗ + A → S + A∗ (2.9)
where S is the sensitizer (initially in an excited state) and A is the activator (in the
ground state). The interaction between the two ions is responsible for the energy
transfer. Transfer via electric dipole-dipole interaction was first described by Förster
and later expanded by Dexter in order to include higher order electromagnetic and
exchange interactions [25], [26]. The dependence of the energy transfer rate on the
S and A separation, R, is R−6 , R−8 and R−10 for the electric dipole-dipole, dipole-
quadrupole and quadrupole-quadrupole mechanisms, respectively. Energy transfer
by exchange interaction requires very short S-A distances, falling off exponentially
with R.
In lanthanide systems, in which there are often several metastable excited states,
cross-relaxation (CR) and UC processes are very common. In the CR mechanism,
part of the high energy excitation of one ion is transferred to the partner which is
typically in the ground state. One of the UC process, as we will see in Section 2.5,
is exactly the reverse.
The temporal evolution of the luminescence from an excited state shows an exponen-
−t
tial behavior in the absence of energy transfer processes, I(t) = I0 exp( τtotal ), where, in
2. Theory 17
general, the total lifetime is due to contributions of both radiative and non-radiative pro-
1 1 1
cesses, τtotal
= τrad
+ τnr . The rate equations model is extremely useful in order to describe
the dynamics when energy transfer processes are involved; the temporal dependence of
the S and A ions population is studied according to a system of coupled differential equa-
tions. This model has been applied in this work, considering an homogeneous impurity
distribution and an average transfer rate, for the fitting of the temporal evolution of the
UC luminescence in order to understand the different UC mechanisms dynamics.
2.3.6 Dimers
When two ions are close enough, they may interact strongly, lose their individual identities,
and become a dimer. As a consequence, the dimer spectroscopic properties are different
from those of the isolated ions. In both exchange (direct overlap of two ions orbitals) and
superexchange (ions coupled through non-magnetic intermediaries) pathways, the inten-
sity of the interaction strongly depends on the distance between the ions in the dimer,
the geometry, and the electronic configuration of the ions. The so-called Goodenough-
Kanamori rules correlate superexchange interactions with the structural properties. The
sign and magnitude of the resulting kinetic superexchange can be predicted by considera-
tion of the symmetry and the electron occupancy of the interacting orbitals on neighboring
ions, in particular, superexchange coupling is very sensitive to the bond angle [27], [28],
[29].
2.4 Raman Spectroscopy

The Raman effect is a light scattering phenomenon by which part of the incident photons
frequency is used to excite vibrational modes of a molecule or solid. Most of the scattered
photons emerge at the same frequency as the incident ones, ω0 , this is elastic or Rayleigh
scattering. On the contrary, inelastic or Raman scattering is a much less probable event
and takes place at new modified frequencies, ω0 ± ωk , where ωk are the frequencies of the
Raman active vibrational modes [22]. Radiation scattered with a frequency lower than
that of the incident light, ω0 − ωk , or Stokes scattering, is more intense than the one at
higher frequency, ω0 + ωk , or anti-Stokes scattering. This light scattering with change of
18 2.4. Raman Spectroscopy
frequency was first observed independently by Raman and Krishnan in liquids [30], [31]
and Mandelstam in quartz [32] in 1928.
In a typical Raman spectrum, the measured frequencies are expressed as the difference
in wavenumbers (cm−1 ) between the incident and scattered photons, regardless of the
excitation wavelength. Rayleigh scattering is a thousand times stronger than Raman
scattering for strong Raman bands. The ratio of the intensities of the anti-Stokes signal
to the Stokes one is related to the thermal population of the excited vibrational levels
and decreases rapidly when the vibrational mode frequency, ωk , increases. Raman spectra
are strongly dependent on temperature [22]. The Raman scattering cross-section for a
semiconductor is given by:
d2 σ ω4
= vV 4s |εs · χ0 · ε0 |hU U + iω (2.10)
dΩdωs c
with
hv|Hint |cihc|Hep |cihc|Hint |vi

|εs · χ0 · ε0 | ≈ (2.11)
(ωg + ωk − ω0 )(ωg − ω0 )
and
~
hU U + iω = (nk + 1)gk (ω) (2.12)
2N ωk
where ωs is the scattered radiation frequency, V is the total volume of the sample, v is
the excited cross-section, c is the light speed, Hint and Hep are the radiation-electron
and electron-phonon interactions, v and c are the intermediate valence band (VB) and
conduction band (CB) states, ωg is the bandgap frequency, ωk are the frequencies of
the Raman active vibrational modes, and ω0 is the frequency of the incident photons.
Moreover, nk is the phonon distribution as a function of temperature, and gk (ω) is the
phonon lorentzian line-shape factor:
1
nk = (2.13)
exp ~ωk
kT
−1
Γk
2π
gk (ω) = (2.14)
Γk 2

(ωk − ω)2 + 2
2. Theory 19
The intensity of Raman scattering is defined by I = P α2 ω04 , where P is the laser

power, and α is the polarizability of the electrons. The basic selection rules for Raman
scattering arises from the change in polarizability. This means that, in contrast to infrared
(IR) absorption, symmetric vibrations give the most intense Raman scattering. In the
case of solid crystalline samples, when radiation interacts with the material, it induces
vibrations through the whole lattice. Longitudinal (L) modes, along the direction of the
radiation propagation, or transverse (T) modes, in a perpendicular direction, occupy a
band of energies in the material. Lower energy lattice modes are called acoustic modes
and labelled LA and TA while the higher energy type are called optic modes, LO and TO
[33].
2.5 Upconversion
In UC processes, the emission of a high energy photon, usually in the visible, takes place
upon the absorption of two or more lower energy photons, in the IR region. These kind
of processes allow the conversion of the incident radiation in high energy light. UC
phenomena was discovered independently by Auzel [34], [35] and Ovsyankin and Feofilov
[36] in 1966. Opposite to second-harmonic generation, UC processes rely on the presence
of real metastable excited states; one intermediate state which acts as a reservoir of
the incident radiation, and at least one higher energy excited state from which emission
occurs. The most relevant UC mechanisms are summarized in Fig. 2.3; the majority of
them involve some combination of absorption and non-radiative energy transfer steps.
Fig. 2.3(a) shows the ground-state absorption/excited-state absorption (GSA/ESA)
mechanism. In the GSA process the active ion is excited to an intermediate level, and it
is then promoted to an upper emitting state by ESA of a second photon. This process is a
single chromophore process and exhibits an immediate decay of the UC luminescence after
the excitation pulse, because both GSA and ESA steps occur during the laser pulse. The
efficiency strongly depends on the oscillator strength of both GSA and ESA processes, as
well as on the resonance of the involved transitions with the excitation energy.
Fig. 2.3(b) describes the energy transfer upconversion (GSA/ETU) process. In this
case, two ions excited by GSA to the intermediate level interact non-radiatively in order
20 2.5. Upconversion
Figure 2.3: Schematic representation of the most important UC mechanisms: GSA/ESA (a), GSA/ETU
(b), cooperative luminescence (c), and cooperative sensitization (d).
to generate one ion in the upper state and the other in the ground state. Since it is a
non-radiative process, ETU can proceed after the pulse, and this is identified as a rise
in the UC luminescence temporal evolution after a short laser pulse excitation. As we
have seen in Section 2.3.5, the efficiency of the ETU mechanism depends on the distance
between the two involved ions. Hence, ETU processes are very sensitive to active ions
concentration.
Fig. 2.3(c) shows the cooperative luminescence. In this case, the involved ions have
no metastable state to emit visible light. However, the simultaneous emission of two IR
photons in two interacting ions give rise to an emitting visible photon of double energy.
This is a very common process in Yb3+ -doped systems [37], [38], [39].
Fig. 2.3(d) describes the cooperative sensitization mechanism in which two excited
sensitizer ions simultaneously transfer their excitation energy to the activator ion. This
ion has no energy level resonant with the excited state of the sensitizer but it presents
an excited state with twice the excitation energy from which emission takes place after
non-radiative energy transfer. Its temporal behavior is expected to be similar to the
GSA/ETU because it also involves an energy transfer step.
It is well known that the number of photons involved in an UC luminescence process
is indicated by the slope of the luminescence intensity versus the power excitation in a
double-logarithmic representation, I ∝ P n . This is only true, however, in the low pump
power regime. Under high power excitation, the power dependence varies with the UC
mechanism nature [40], [41]. Hence, both GSA/ESA and GSA/ETU shows a quadratic
2. Theory 21
power dependence under low power excitation conditions, and linear dependence at high
power. The efficiency of the different UC processes depends on the number of involved
optically active centers. Cooperative transitions involving three ions are about 3-4 orders
of magnitude weaker than transitions between one or two ions states [42].
2.6 Rare-earth ions

Lanthanides or RE elements are those situated in the sixth period of the periodic table
after the lanthanum, from cerium with the outer configuration 5s2 5p6 4f1 5d1 6s2 , to ytter-
bium with configuration 5s2 5p6 4f14 6s2 [20], [18]. These atoms are usually incorporated in
crystals as divalent or trivalent cations. In RE3+ ions, 5d, 6s and some 4f electrons are re-
moved; hence, trivalent ions deal with transitions between sub-levels of the 4fn electronic
configuration covering an energy range from NIR to deep UV [43]. The optically active
4f electrons are shielded by the 5s and 5p outer electrons. Because of this shielding effect,
the valence electrons are weakly affected by the ligand ions in crystals. This situation
corresponds to the weak crystal field case (see Section 2.3.2). The effect of the crystal field
is to produce a slight shift in the energy of the 2S+1 LJ RE3+ states and to cause additional
splitting. This implies that the main features of RE3+ ions spectra are similar comparing
isolated ions to different hosts. On the contrary, RE2+ ions interact with the crystalline
environment due to the presence of 5d electrons, and show f ↔ d intraconfigurational
transitions, strongly dependent on the host crystal symmetry. These f ↔ d transitions
are also observed in some RE3+ ions.
The interpretation of the RE3+ ions optical spectra is based on the so-called Dieke
2S+1
diagram (Fig. 2.4). Dieke et al. determined the energy of the LJ states for the RE3+
ions in LaCl3 up to 40000 cm−1 [44]. Wegh et al. extended the Dieke diagram from 40000
to about 70000 cm−1 for various lanthanide ions [43]. The width of each state indicates
the magnitude of the crystal-field splitting, it is smaller than 1000 cm−1 in all cases. This
value is in general lower than the spin-orbit splitting. The free ion energy levels are given
by the center of gravity of each multiplet. The same spin-orbit terms in an inverse energy
order for the 4fn and 4f14−n configurations is evidenced in Dieke diagram. Besides, a
comparison of the energy levels of Ce3+ and Yb3+ shows clearly the spin-orbit parameter
22 2.6. Rare-earth ions
Figure 2.4: Dieke diagram of RE3+ ions in LaCl3 [44]. Semicircles indicate the emitting levels.
2. Theory 23
increase for larger atomic number.

Since electron-phonon coupling effects are weak, optical transitions between 4fn states
in RE3+ are characterized by sharp lines. These transitions occur between states of the
same parity and then they are expected to occur by a magnetic dipole process. However, a
mixing of 4fn states with opposite-parity states of the 4fn−1 -5d configuration can take place
by coupling with odd vibrational modes, allowing a forced electric dipole component to
the transition. Spin is not a good quantum number to consider in f-f transitions because
of the spin-orbit coupling. Some general selection rules for electric dipole transitions
between RE3+ 4fn states obtained from the Judd-Ofelt theory are |∆J| ≤ 6 with J =
0 ↔ J 0 = 0 forbidden, and J = 0 ↔ odd J 0 much weaker than J = 0 ↔ even J 0 [20].
Interconfigurational 4fn ↔ 4fn−1 5d transitions are parity allowed and usually appear in
the UV region for RE3+ ions. Their energy strongly depends on the local environment
because 5d levels are not so efficiently screened.
2.7 Transition metal ions

TM ions (iron group) are formed from the elements situated in the fourth period of
the periodic table in which the 3d shell is being filled. In ionic solids, such elements
lose the outer 4s electrons and some 3d electrons, and their electronic configurations are
1s2 2s2 2p6 3s2 3p6 3dn , where n < 10 [20]. Since 3d electrons, responsible for the optical
transitions, interact strongly with electric fields of nearby ions, the crystal-field effects
are, in general, more important than for RE3+ ions. Besides, TM ions show strong
electron-lattice coupling; as a result, the spectra can present both broad (S > 0) and
sharp (S ∼ 0.1) bands [18]. According to the strong crystal field scheme, and taking
into account that the electron-electron and crystal-field interactions dominate over the
spin-orbit term, the electronic levels splitting of the TM ion can be obtained according
to each complex symmetry [20]. In an octahedral crystal field, the five d levels are split
into a triply-degenerate level t2g and a doubly-degenerate level eg ; the strength of the
crystal field responsible for this splitting, ∆, is called 10Dq. The energy of the transitions
observed experimentally in the TM spectra can be described by using ∆=10Dq and the
so-called Racah parameters B and C [45].
24 2.7. Transition metal ions
Figure 2.5: Tanabe-Sugano diagram for a d3 ion in an octahedral crystal field or a d7 ion in a tetrahedral
crystal field (left), and Tanabe-Sugano diagram for a d5 ion (right). The vertical lines indicate the crystal
field strength for Co2+ , Cr3+ and Mn2+ ions, in the lattices studied in this work; ZnO, Gd3 Ga5 O12 and
LaMgAl11 O19 , respectively.
Sugano and Tanabe calculated the energy of the states deriving from dn ions (2<
n <8) as a function of the octahedral crystal field. These calculations are represented in
the Tanabe-Sugano diagrams which show the energy of the states in units of B (E/B )
compared to the ground state energy, as a function of the ratio between the crystal-field
strength and the interelectronic interaction, measured in units of ∆/B, for a given C/B
ratio [46]. The energy level scheme of a 3d10−n ion in a tetrahedral crystal field is the same
as in an octahedral crystal field but with negative 10Dq, hence, it has the same pattern
as the splitting of a 3dn system in an octahedral crystal field. The relationship between
the crystal-field strenghts is given by 10Dq(octahedral)= − 49 10Dq(tetrahedral). Figure
2.5 shows the Tanabe-Sugano diagrams for d3 and d5 systems in an octahedral crystal
field. These diagrams are appropriate to describe the TM ions that have been studied
in this work: Cr3+ (3d3 system) in an octahedral crystal field and Co2+ (3d7 system) in
a tetrahedral crystal field (Fig. 2.5(a)), and Mn2+ (3d5 system) in a tetrahedral crystal
2. Theory 25
field (Fig. 2.5(b)). Tanabe-Sugano diagrams allow to predict if an optical band of a TM

ion is narrow or broad from the slope of the involved states. In general, the magnitude of
the slope is proportional to the electron-lattice coupling constant. Therefore, the states
with energies independent of the crystal field give rise to narrow bands while broad bands
are associated with transitions involving large slope energy levels. Despite the fact that
non-radiative processes are more important in TM than in RE systems, UC luminescence
processes have also been observed in different TM ions such as Pb+ , Ti2+ , Mn2+ , Ni2+ or
Cr3+ [47], [48], [49], [50].
2.8 Semiconductor nanostructures
The bandgap of a semiconductor is defined as the energy difference between the highest
energy VB states and the lowest energy CB states. If an electron is excited from the
VB to the CB leaving a hole in the VB, the electron-hole (e-h) pair may form a bound
state through Coulomb interactions. This bound state (the so-called Wannier-Mott or
free exciton) has an energy slightly lower than the bandgap energy. The Bohr radius of
the exciton, aB , is given by the following equation:
4πε0 ε∞ ~2 1 1
aB = = a 0 ε ∞ (2.15)
m0 e2 µ∗ µ∗
where ε0 is the vacuum dielectric constant, ε∞ is the high frequency relative dielectric
constant of the medium, m0 is the mass of the free electron, µ∗ , with 1
µ∗
= 1
m∗e
+ 1
m∗h
,
is the effective reduced e-h mass (m∗e and m∗h are the effective electron and hole masses),
and a0 is the Bohr radius of a hydrogen atom (0.529 Å). Since the effective masses are
significantly smaller than m0 , and ε∞ is much larger than 1, the resulting values for aB
are much larger than a0 , in the range 1-10 nm for the common semiconductors. The Bohr
radius of the exciton is 2.6 nm for bulk CdS in the ZB structure [51] and 0.9 nm for bulk
W-ZnO [52]. Semiconductor nanoparticles exhibit quantum size or quantum confinement
effects when the dimensions of the system approaches the Bohr radius of the exciton, as a
result of spatial confinement of the electrons and holes in the nanoparticles; in these cases,
they are also called QDs [53], [54]. Two manifestations of these quantum size effects are
the increase in the bandgap with decreasing particle size, and the transition from energy
bands to discrete energy levels. The chemical bonds contraction at the surface due to
coordination deficiency and the rise of the surface-to-volume ratio have been shown to be
responsible for these changes at the nanoscale [55].
Different approaches have been proposed to explain the bandgap expansion of semi-
conductor nanocrystals [55]. The widely accepted model was initially proposed by Efros
[56] and further extended by Brus [53], [57] to include the term of Coulomb interaction of
an e-h pair. The confinement effect on the bandgap, Eg , is then expressed as a function
of the semiconductor nanoparticle radius, R:
π 2 ~2 1 1.8e2
Eg (R) = Eg + − (2.16)
2m0 R2 µ∗ 4πε0 ε∞ R
where the first term, Eg , represents the bandgap of the bulk semiconductor, the second
term represents the spatial confinement of electrons in the CB and holes in the VB by
the potential barrier of the surface or a mono-potential well of the quantum box and
has a 1/R2 dependence, and the third term represents the Coulomb energy with a 1/R
dependence. Generally, the second term is more important than the third one, leading to
an increase in the QDs energy bandgap when reducing the particle size.
2.9 Interband absorption in semiconductors

2.9.1 Introduction
Semiconductors absorption spectra in the visible region are due to interband transitions,
in which an electron is excited from the VB to the CB. A continuous absorption spectrum
starting from the low energy threshold at the energy bandgap, Eg , is observed for bulk
semiconductors. A step-like behavior is expected for quantum wells while a series of
delta functions is predicted for QDs [17]. Although the joint density of states changes
when the QDs size is reduced, the only size effect that has been considered in this work
is the blue-shift of the absorption edge with decreasing particle size. Therefore, the
equations describing interband transitions in bulk semiconductors have been used, and
are shown in the next two sections. The interband absorption rate depends on the band
2. Theory 27
structure of the solid. Considering the relative positions of the CB minimum and the VB
maximum in the Brillouin zone, the bandgap may be direct or indirect. Conservation
of momentum implies that absorption processes are represented by vertical lines on E-k
diagrams. Hence, if the bandgap is indirect, the transition must involve a phonon to
satisfy momentum conservation [17].
Figure 2.6: Interband transitions in semiconductors. Direct bandgap (a) and indirect bandgap (b).
2.9.2 Band edge absorption in direct gap semiconductors
In a direct gap semiconductor, both the VB maximum and the CB minimum occur at the
same value of k (see Fig. 2.6(a)). Assuming a simplified parabolic form for the VB and
CB, the conservation of energy during a transition requires that:
~2 k 2
~ω = Eg + (2.17)
2µ
where µ is the reduced e-h mass. The joint e-h density of states, g(~ω), is then given by:
g(~ω) = 0 For ~ω < Eg (2.18)

3/2
1 2µ
g(~ω) = 2
(~ω − Eg )1/2 For ~ω ≥ Eg
2π ~2
Fermi’s Golden Rule tells us that the absorption rate for a dipole-allowed interband
transition is proportional to the joint density of states. Hence, the following behavior is
expected for the frequency dependence of the absorption coefficient, α(~ω) [17]:
α(~ω) = 0 For ~ω < Eg (2.19)
α(~ω) ∝ (~ω − Eg )1/2 For ~ω ≥ Eg
According to eq. (2.19), a straight line behavior is obtained plotting α2 against the
photon energy, ~ω, in the spectral region close to the bandgap. The bandgap is the point
at which the absorption goes to zero.
2.9.3 Band edge absorption in indirect gap semiconductors
Indirect gap materials have their CB minimum away from the Brillouin zone center, as
shown schematically in fig. 2.6(b). A photon can excite an electron from the VB to the CB
involving a phonon to conserve momentum, k. Conservation of energy and momentum
requires that:
~ω = ECV ± ~Ω (2.20)
kV − kC = ±Q (2.21)
where ECV is the difference in energy between VB and CB, ~Ω is the phonon energy, kV
and kC are the respective wave vectors of the involved states in VB and CB, and Q is the
phonon wave vector. The ± sign represents the phonon absorption or emission. Indirect
transitions are second-order processes, the transition rate is therefore much smaller than
for direct absorption. The frequency dependence of the indirect absorption coefficient in
this case is given by:
αi (~ω) ∝ (~ω − Eg ∓ ~Ω)2 (2.22)
If α1/2 is plotted against ~ω, the indirect gap semiconductor data fits well to a straight
line. The experimental data also shows a tail caused by higher frequency phonons or
multiphonon absorption.
It must be pointed out that since the frequency dependence is different for direct and
indirect gap semiconductors, it provides a very convenient way to determine the nature
of the bandgap.
2. Theory 29
2.10 Structural phase transitions. Equations of state
The crystalline structure of a solid may undergo modifications when pressure or temper-
ature conditions are changed. The different phases of a given material together with its
P − T stability points compose the phase diagram of the material. From a thermody-
namic point of view, a given phase is stable when the Gibbs free energy is minimum.
First-order phase transitions exhibit a discontinuity in the first derivative of the Gibbs
free energy. A diminution of the volume or the entropy is experimentally observed at
the transition point. Transitions showing hysteresis in both pressure and temperature are
first-order transitions. Second-order phase transitions are continuous in the first deriva-
tive but exhibit discontinuity in a second derivative of the free energy, they take place at
well-defined temperature and pressure, and they present no hysteresis [58]. It is impor-
tant to point out that pressure-induced phase transitions lead to transformations to more
compact structures, increasing the coordination number and involving volume reduction.
Density increases and, therefore, they are usually first-order phase transitions.
The appearance of a structural phase transition may have important consequences

not only in the structural characterization measurements, but also in the optical proper-
ties. The absorption or emission spectra of an optically active ion change drastically if
the new phase involves modifications of the coordination number or bond distances. Be-
sides, changes in the bandgap absorption and Raman spectra of semiconductor materials,
associated to the transformation into a new crystalline structure are also very common.
An equation of state provides a relationship between two or more state variables, such
as temperature, pressure or volume. For a given phase, temperature, volume, and pressure
are not independent quantities; they are connected by a relationship of the general form
f (P, V, T ) = 0. For first-order phase transitions, each structural phase in the P − T
diagram has a characteristic equation of state.
Considering an isothermal relation between P and V , the simplest reasonable model

∂P

for an equation of state would be with a constant bulk modulus, B0 = −V ∂V T
. A more
sophisticated equation of state was derived by Murnaghan assuming that isothermal B is a
linear function of P , B = B0 + B00 · P . The bulk modulus pressure derivative, B 0 = ∂B

∂P T
,
30 2.10. Structural phase transitions. Equations of state
is considered constant, B 0 = B00 , and therefore [59]:
−1/B00
B0

V (P ) = V0 1 + 0P (2.23)
B0
V0 being the equilibrium volume, V (P = 0). This equation is appropriate for compressions
lower than 15%, however, for higher compressions more elaborated equations of state are
needed, such as Birch-Murnaghan equation [60] or Rose-Vinet equation [61].
Bibliography
[1] G.L. Hornyak, J. Dutta, H.F. Tibbals, and A.K. Rao, editors. Introduction to
Nanoscience. CRC Press, Taylor and Francis group, Florida, 2008.
[2] A. Trave, F. Buda, and A. Fasolino. Band-Gap Engineering by III-V Infill in Sodalite.
Phys. Rev. Lett., 77: 5405–5408, 1996.
[3] J.R. Agger, M.W. Anderson, M.E. Pemble, O. Terasaki, and Y. Nozue. Growth of
Quantum-Confined Indium Phosphide inside MCM-41. J. Phys. Chem. B, 102: 3345–
3353, 1998.
[4] F. Vetrone, J.C. Boyer, J.A. Capobianco, A. Speghini, and M. Bettinelli. A spec-
troscopic investigation of trivalent lanthanide doped Y2 O3 nanocrystals. Nanotech.,
15: 75–81, 2004.

[6] E.L. Wolf, editor. Nanophysics and Nanotechnology. Wiley-VCH, Weinheim, Ger-
many, 2004.
[7] M. Stupca, M. Alsalhi, T. Al Saud, A. Almuhanna, and M.H. Nayfeh. Enhancement

of polycrystalline silicon solar cells using ultrathin films of silicon nanoparticle. Appl.
Phys. Lett., 91: 063107, 2007.
31
32 BIBLIOGRAPHY
[9] R. Elghanian, J.J. Storhoff, R.C. Mucic, R.L. Letsinger, and C.A. Mirkin. Selective
Colorimetric Detection of Polynucleotides Based on the Distance-Dependent Optical
Properties of Gold Nanoparticles. Science, 277: 1078–1081, 1997.
[10] W.C.W. Chan and S. Nie. Quantum Dot Bioconjugates for Ultrasensitive Nonisotopic
Detection. Science, 281: 2016–2018, 1998.
Lett., 8: 3834–3838, 2008.
[12] F. Vetrone and J.A. Capobianco. Lanthanide-doped fluoride nanoparticles: lumi-

nescence, upconversion, and biological applications. Int. J. Nanotech., 5: 1306–1339,
2008.
[13] C. Vancaeyzeele, O. Ornatsky, V. Baranov, L. Shen, A. Abdelrahman, and M.A.

Winnik. Lanthanide-Containing Polymer Nanoparticles for Biological Tagging Appli-
cations: Nonspecific Endocytosis and Cell Adhesion. J. Am. Chem. Soc., 129: 13653–
13660, 2007.
[14] P.V. Kamat. Meeting the Clean Energy Demand: Nanostructure Architectures for
Solar Energy Conversion. J. Phys. Chem. C, 111: 2834–2860, 2007.
[15] A. Shalav, B.S. Richards, and T. Trupke. Application of NaYF4 : Er3+ up-converting
phosphors for enhanced near-infrared silicon solar cell response. Appl. Phys. Lett.,
86: 013505, 2005.
[16] L. Aarts, B.M. van der Ende, and A. Meijerink. Downconversion for solar cells in
NaYF4 : Er, Yb. J. Appl. Phys., 106: 023522, 2009.
[17] M. Fox, editor. Optical Properties of Solids. Oxford University Press, Oxford, 2001.
BIBLIOGRAPHY 33
[21] R. Valiente, F. Rodrı́guez, J. González, H.U. Güdel, R. Martı́n-Rodrı́guez, L. Nataf,

M.N. Sanz-Ortiz, and K. Krämer. High pressure optical spectroscopy of Ce3+ -doped
Cs2 NaLuCl6 . Chem. Phys. Lett., 481: 149–151, 2009.
[24] J.M.F. van Dijk and M.F.H. Schuurmans. On the nonradiative and radiative decay
rates and a modified exponential energy gap law for 4f−4f transitions in rare-earth
ions. J. Chem. Phys., 78: 5317–5323, 1983.
[25] T. Förster. Ann. der Physik, 2: 55, 1948.
[26] D.L. Dexter. J. Chem. Phys., 21: 836, 1953.
[27] J.B. Goodenough, editor. Magnetism and the Chemical Bond. Interscience-Wiley,
New York, 1963.
[28] J. Kanamori. J. Phys. Chem. Solids, 10: 87, 1959.
[29] H. Weihe and H.U. Güdel. Quantitative Interpretation of the Goodenough-Kanamori

Rules: A Critical Analysis. Inorg. Chem., 36: 3632–3639, 1997.
[30] C.V. Raman and K.S. Krishnan. A New Type of Secondary Radiation. Nature,
121: 501–502, 1928.
[31] C.V. Raman. A Change of Wave-length in Light Scattering. Nature, 121: 619–619,
1928.
34 BIBLIOGRAPHY
[32] G.S. Landsberg and L.J. Mandeltsam. Eine neue Erscheinung bei der Lichtzerstreu-
ung in Krystallen. Naturwissenschaften, 16: 557–558, 1928.
[33] W.E. Smith and G. Dent, editors. Modern Raman Spectroscopy - A Practical Ap-
proach. John Wiley and Sons, England, 2005.
[34] F. Auzel. Acad. Sci., 262: 1016, 1966.
[35] F. Auzel. Acad. Sci., 263 B: 765, 1966.
[36] V.V. Ovsyankin and P.P. Feofilov. Jetp. Lett., 3: 322, 1966.

Lett., 25: 1710–1712, 1970.

J. Chem. Phys., 98: 1768–1775, 1993.
Phys. Rev. B, 61: 3337–3346, 2000.
2005.
[42] F. Auzel. Upconversion and Anti-Stokes Processes with f and d ions in Solids. Chem.
Rev., 104: 139–173, 2004.
[43] R.T. Wegh, A. Meijerink, R.J. Lamminmäki, and J. Hölsä. Extending Dieke’s dia-
gram. J. Lumin., 87-89: 1002–1004, 2000.
[44] G.H. Dieke, editor. Spectra and Energy Levels of Rare Earth Ions. Wiley Interscience,
New York, 1968.
BIBLIOGRAPHY 35
[45] S. Sugano and Y. Tanabe. J. Phys. Soc. Jpn., 9: 753, 1954.
[46] S. Sugano, Y. Tanabe, and H. Kamimura, editors. Multiplets of transition metal ions
in solids. Academic Press, New York, 1970.
[47] J.M. Spaeth, R.H. Bartram, M. Rac, and M. Fockele. Upconversion by excited state
absorption of Pb+ (l) centres in alkaline-earth fluorides. J. Phys.: Condens. Matter,
3: 5013–5022, 1991.
[49] S. Garcı́a-Revilla, P. Gerner, H.U. Güdel, and R. Valiente. Yb3+ -sensitized visi-
ble Ni2+ photon upconversion in codoped CsCdBr3 and CsMgBr3 . Phys. Rev. B,
72: 125111, 2005.
2002.
[51] Y. Kayanuma. Quantum-size effects of interacting electrons and holes in semicon-

ductor microcrystals with spherical shape. Phys. Rev. B, 38: 9797–9805, 1988.
[53] L.E. Brus. Electron-electron and electron-hole interactions in small semiconductor

crystallites: The size dependence of the lowest excited electronic state. J. Chem.
Phys., 80: 4403–4409, 1984.
[54] P.F. Trwoga, A.J. Kenyon, and C.W. Pitt. Modeling the contribution of quantum
confinement to luminescence from silicon nanoclusters. J. Appl. Phys., 83: 3789–3794,
1998.
[55] C.Q. Sun, T.P. Chen, B.K. Tay, S. Li, H. Huang, Y..B Zhang, L.K Pan, S.P. Lau,
and X.W. Sun. An extended quantum confinement theory: surface-coordination
36 BIBLIOGRAPHY
imperfection modifies the entire band structure of a nanosolid. J. Phys. D: Appl.

Phys., 34: 3470–3479, 2001.
[56] Al.L. Efros and A.L. Efros. Interband absorption of light in a semiconductor sphere.
Sov. Phys. Semicond., 16: 772–775, 1982.
[57] L.E. Brus. On the development of bulk optical properties in small semiconductor
crystallites. J. Lumin., 31-32: 381–384, 1984.
[58] R. Ehrenfest. Proc. Acad. Sci. Amsterdam, 36: 153–157, 1933.
[59] F.D. Murnaghan. The Compressibility of Media under Extreme Pressures. P. Natl.
Acad. Sci., 30: 244–247, 1944.
[60] F. Birch. Finite Elastic Strain of Cubic Crystals. Phys. Rev., 71: 809–824, 1947.
[61] P. Vinet, J. Ferrante, J.R. Smith, and J.H. Rose. A universal equation of state for
solids. J. Phys. C: Solid State Phys., 19: L467–L473, 1986.
Chapter 3
Synthesis
3.1 Introduction
In this chapter, the different synthesis methods used for the preparation of the samples
are described. Er3+ and Yb3+ co-doped Y2 O3 and NaYF4 nanoparticles, as well as CdS
nanopowders were prepared in a planetary ball mill at the Solid State Physics laboratory of
the University of Cantabria. RE-doped Y2 O3 nanocrystals were also synthesized following
a combustion reaction at the Inorganic Chemistry laboratory of our university. Tb3+ -
Yb3+ and Eu3+ -Yb3+ co-doped ternary oxides GGG and YAG nanocrystalline samples,
and Cr3+ -doped GGG nanoparticles were grown during a two month stay within Prof.
Marco Bettinelli’s group at the Biotecnology Department, University of Verona. All of
them were obtained using Pechini’s method [1]. Some of these samples were coated with
SiO2 using Stöber method [2]. As another result of this collaboration, microcrystalline
powders of LMA doped with Mn2+ and Yb3+ were also synthesized in one of Prof. Marco
Bettinelli’s laboratory. Colloidal Zn1−x Cox O nanocrystals were prepared at the Chemistry
Department, University of Washington, thanks to a collaboration with Prof. Daniel R.
Gamelin during a three month stay in his group.
3.2 Mechano-chemical processes in a planetary ball

mill
A wide variety of processes such as crystallization of amorphous alloys, particle size re-
duction or mechano-chemical reactions can take place in a planetary ball mill [3]. In all
37
cases, mechanical ball milling consists of repeated fracture, mixing and welding of pow-
der particles (less than 10 µm in size). When the mill rotates in one direction the vials
rotate in the opposite one, and there are collisions among the powders, the balls and the
vials walls [4]. The occurrence of these mechano-chemical reactions is attributed to the
heat generated in the milling process, favored by the large area of contact between the
solids [5]. At the point of collision, the solids deform and even melt, forming hot points
where the molecules can reach very high vibrational excitation leading to bond breaking.
Related to the final nanoparticles size, it is important to point out the existence of a
steady-state grain size during high-energy ball-milling. If the initial size is bigger than
the steady-state size, the crystallite size decreases during the milling, but smaller particles
show an increase in size reaching finally such steady-state. This can be explained by an
equilibrium between fracturing and crystal growth processes during the milling [6].
Figure 3.1: Planetary ball mill used for the synthesis of Y2 O3 , NaYF4 and CdS nanoparticles at the
Solid State Physics laboratory, University of Cantabria.
The high-energy ball-milling synthesis was carried out using a planetary ball mill
(Retsch PM 400/2) (Fig. 3.1). The most important parameters to take into account in
these processes are: Vials and balls material, balls size, atmosphere under which synthesis
is performed (Air or argon), total milling time, milling and pause times, angular velocity
(rpm), ball to sample mass ratio and the addition of a diluent. Changes in every parameter
are decisive not only for the obtention of the desired phase but also for the final size of the
nanocrystals. Searching for the optimum conditions for each synthesis process involved
an arduous and systematic work. In some cases, the ball mill was also used in order to
3. Synthesis 39
homogenize the size of nanoparticles prepared using other synthesis methods.
Y2 O3 : Er3+ , Yb3+ nanoparticles were obtained in a planetary ball mill by intimately

mixing stoichiometric quantities of yttrium, erbium and ytterbium oxides according to
the reaction:
(1 − A − B)Y2 O3 + AEr2 O3 + BYb2 O3 → Y2 O3 : A%Er3+ B%Yb3+ (3.1)
The following parameters were used:
• ZrO2 vials and balls, with ball diameter of 10 mm. Other materials such as steal
had been previously tried, but they contaminated the samples.
• Ball to powder mass ratio of 10:1, with a total amount of 6 g.
• Air atmosphere and 200 rpm.
• 5 minute stop every 30 minutes of milling. The dependence of the particle size on
the total milling time was studied in order to determine the most advantageous
conditions.
Different procedures have been previously described in the literature for the synthesis of
sodium yttrium fluoride, NaYF4 , such as heating the mixture NaF + YF3 at 800 ◦ C under
HF gas [7] or a hydrothermal method that requires complicated apparatus [8]. These re-
strictions in fluoride synthesis have limited their development in comparison, for instance,
with oxides. Fluorides have also been prepared by mechano-chemical synthesis [9]. This
fact, and the difficulty in synthesizing the pure hexagonal phase by other methods, led
us to prepare hexagonal NaYF4 : Er3+ , Yb3+ nanocrystals using the planetary ball mill
starting from stoichiometric quantities of fluorides according to the reaction:
NaF + (1 − A − B)YF3 + AErF3 + BYbF3 → NaYF4 : A%Er3+ , B%Yb3+ (3.2)

Bearing in mind the necessity of using parameters which increase collisions energy
during the milling [9], it is convenient to increase the ball to powder mass ratio or the
angular velocity. A systematic study to obtain the optimal synthesis conditions was
accomplished; different grinding times, velocity and mass ratio were tried. The final
parameters established as the most convenient for the synthesis of hexagonal NaYF4
were:
• ZrO2 vials and 20 balls.
• Ball to sample mass ratio of 30:1.
• Argon atmosphere and 300 rpm.
• 2 hours stop every grinding hour.
Different total milling times were used in order to get the pure hexagonal phase.
Changes in any of the former parameters are rather critical for the obtention of the desired
hexagonal phase. For example, if a ball to powder mass ratio of 20:1 is used, cubic phase
is obtained after 5 hours milling. If a 5 minute pause is set, peaks corresponding to cubic
phase are more intense than hexagonal peaks after 8 hours milling.
3.2.3 Pure and Yb3+ -doped CdS
Pure and Yb3+ -doped CdS nanoparticles were prepared according to the following mechano-
chemical reaction in a planetary ball mill [10], [11]:
CdCl2 + YbCl3 + Na2 S + 10NaCl → CdS:Yb3+ + 12NaCl (3.3)
The starting materials were CdCl2 , YbCl3 , Na2 S and NaCl. Milling was performed
under high purity argon atmosphere using hardened steel vials and balls, a ball to sample
mass ratio of 20:1 and 300 rpm. NaCl was used as diluents in order to prevent nanopar-
ticles aggregation and has little effect on the obtained particles size. The particle size
of the product depends on the size of the starting reactants; smaller Na2 S particles give
rise to smaller CdS nanocrystals [12]. In this sense, the mixture Na2 S + 10NaCl was ini-
tially milled for 1 hour. Then, CdCl2 was added and the mixture was milled for another
3. Synthesis 41
hour. The as-obtained powders were washed with distilled water, filtered and dried under
vacuum in order to remove NaCl contents.
3.3 Combustion reaction

Y2 O3 nanocrystals doped with different Er3+ and Yb3+ concentrations were synthesized
following the glycine-nitrate solution combustion reaction:
3 5 25
3M(NO3 )3 + 5NH2 CH2 COOH + 9O2 → M2 O3 + N2 + 9NO2 + 10CO2 + H2 O (3.4)
2 2 2
where M = Y3+ , Er3+ or Yb3+ [13], [14], [15]. The experimental procedure was as follows:
yttrium, ytterbium, erbium nitrates and glycine were mixed in an appropriate ratio and
dissolved in distilled water to form the precursor solution. This solution was concentrated
by heating in a quartz crucible or in a pyrex beaker on a heating plate. When excess
water was evaporated, a spontaneous ignition took place and the combustion reaction
was finished originating the oxide nanoparticles (fig. 3.2).
Figure 3.2: Different sequences during the combustion synthesis of RE-doped Y2 O3 nanoparticles at
the Inorganic Chemistry laboratory, University of Cantabria.
The combustion temperature has a great influence on the final particle size, and this
reaction temperature can be controlled by changing the glycine-to-nitrates ratio (G/N).
For the preparation of RE oxides, it has been reported that lower temperature leads
to smaller size for the obtained nanoparticles [15]. In this sense, a G/N lower than
stoichiometric, G/N=1, was used in order to get as small nanoparticles as possible.
In the described process, Y2 O3 nanocrystals adsorb part of the H2 O and CO2 molecules
produced during the reaction. The presence of these groups on the nanoparticles surface
42 3.4. Pechini’s method (sol-gel)
yields higher energy phonons which make multiphonon relaxation much more probable,
and are responsible for the reduction in the luminescence (see Section 5.2.1). Different
heat treatments were carried out in order to reduce the amount of OH− and CO2−
3 ions
which were analyzed by IR absorption spectroscopy. All samples were fired at different
temperatures for different periods of time. It must be considered that longer treatments at
higher temperatures reduce surface contamination, but they can also induce nanoparticles
aggregation with the resulting increase in size.
3.4 Pechini’s method (sol-gel)

GGG and YAG nanocrystalline powders co-doped with Tb3+ or Eu3+ and Yb3+ , and
Cr3+ -doped GGG nanoparticles have been prepared using the sol-gel Pechini’s method as
described in Ref. [16]. Among the sol-gel procedures, the preparation of oxides using the
formation of an intermediate precursor polyester was proposed by Pechini in 1960’s [1].
This technique was initially proposed to prepare niobates and titanates ceramic materials
or thin films. Pechini’s method is based on the ability of some α-hydroxycarboxylic acids,
such as citric acid (CA), to form chelates between the metal ions (Gd, Ga, Y or Al in
our case) and the acid. When this complex is mixed with some polyhydroxyl alcohol,
usually polyethylene glycol (PEG), and is heated, the polymerization process takes place.
Finally, the polymeric matrix is thermally decomposed to remove the organic groups and
form the stoichiometric oxide phase [17].
In our case, the synthesis of ternary oxides according to this method was as follows:
First, the precursor solution was prepared by dissolving the nitrates in distilled water
and adding a suitable amount of CA, according to the ratio CA moles = 2 × Nitrates
moles. The starting materials were reagent grade Gd(NO3 )3 , Ga(NO3 )3 and Ln(NO3 )3 or
Cr(NO3 )3 for the GGG samples and Y(NO3 )3 , Al(NO3 )3 and Ln(NO3 )3 in the case of the
YAG samples. Then, the solution was heated at 90 ◦ C under stirring until a transparent
solution was obtained. At this step, molecules between metal ions and acid are formed.
Afterwards, PEG was added to the solution according to PEG moles = CA moles / 20
and it was stirred for 15 minutes so that polymerization could take place. Finally, the
obtained sol was heated at 90 ◦ C for 24 hours in order to form the gel, an amorphous
3. Synthesis 43
mixture of oxides, and this gel was fired at 800 ◦ C for 16 hours to induce its crystallization
and to form GGG and YAG nanoparticles (Fig. 3.3).
Figure 3.3: Sol, gel, and powder steps during the GGG and YAG nanoparticles synthesis by Pechini’s
method at the Biotecnology Department, University of Verona.
3.5 Nanoparticles coating

Recently, the fabrication of core-shell structures has attracted increasing attention. Ini-
tially, the idea of surface modification was applied in semiconductor nanoparticles but it
has also been extended to inorganic luminescent systems [18], [19], [20]. The formation
of core-shell structures is an effective way to improve luminescence efficiency since the
shell protects the core from interaction with the surrounding medium [21]. Moreover, in
some cases such as SiO2 , the coating allows the nanocrystals to be redispersible in organic
solvents.
A system of chemical reactions to control the growth of spherical silica particles with
sizes in the range 50-200 nm was developed by Stöber in 1968 [22]. A procedure to
incorporate radioactive isotopes into silica particles was also established by Flachsbart and
Stöber [2]. This technique allows the control of the spherical shape for the nanoparticles,
and it was used for the SiO2 coating of Eu3+ doped YAG and Y2 O3 [20], [23].
In this work, SiO2 -coated nanoparticles were obtained from the above Y2 O3 , GGG
and YAG samples using the Stöber method [23], [2]. The experimental details were as
follows: a nanopowder sample (0.153 g) was mixed with 60 ml of isopropyl alcohol in a
100 ml flask. The nanoparticles were dispersed with ultrasonic oscillation for 45 minutes.
Then, 4.5 ml of deionized water and 6 ml of hydrous ammonia were gradually added under
stirring at 50 ◦ C for 10 minutes. After that, 145 µl of tetraethyl orthosilicate (TEOS) was
44 3.6. Precipitation method
added and the solution was stirred at 50 ◦ C for 1 hour. Finally, the reaction mixture was
centrifuged at 4000 rpm for 10 minutes. The resulting nanoparticles were washed three
times by centrifuging in deionized water at 4000 rpm for 10 minutes and dried at 90 ◦ C
for 20 hours.
3.6 Precipitation method

Different procedures have been reported for the synthesis of LMA, such as solid state
reactions [24], a Pechini sol-gel procedure [25], combustion synthesis [26] or a precipitation
method [27]. All these methods are known to produce LMA powders with grain size in
the range of micrometers. In this work, several attempts to prepare LMA host material
by intimately mixing the starting oxides, La2 O3 + 2MgO + 11Al2 O3 , and heating up
to 1400 ◦ C were performed, but a mix of the initial oxides or an amorphous phase was
obtained. Another problem related to the synthesis of this system was the presence of
Al2 O3 as a non-desired impurity. Combustion synthesis was supposed to give rise to single
phase LMA since it requires lower temperatures. However, our experience is that after a
combustion reaction and a heat treatment at 500 ◦ C, an amorphous phase was achieved.
Firing the samples at 1200 ◦ C for more than ten hours was needed in order to obtain
crystalline LMA by combustion. LMA synthesis following the precipitation method was
also performed.
Both methods, combustion and precipitation, led to the formation of crystalline LMA
with traces of Al2 O3 impurities; but, as we will see later (see Section 5.3.1), luminescence
properties are much more interesting for precipitation samples. Therefore, the synthe-
sis of microcrystalline powders of LMA co-doped with different concentrations of Mn2+
and Yb3+ was developed in this work following the precipitation method described in
Ref. [27]. Stoichiometric quantities of La(NO3 )3 5H2 O, Al(NO3 )3 9H2 O, Mg(NO3 )2 6H2 O,
Mn(NO3 )2 H2 O and Yb(NO3 )3 5H2 O were dissolved in distilled water under stirring. Then,
NH4 OH was added dropwise until a pH of 8.5 was reached and precipitation as hydroxides
occurred. The obtained hydroxides were centrifuged at 3000 rpm for 5 minutes, washed
with distilled water, and heated at 90 ◦ C for 20 hours. A treatment at 700 ◦ C for 2 hours
and 1500 ◦ C for 1 hour was carried out to obtain the material in its final form.
3. Synthesis 45
3.7 Colloidal semiconductor nanocrystals
The optical and magnetic properties of wide-gap diluted magnetic semiconductors (DMSs)
are attracting broad interest in both fundamental and applied research areas [28]. Two
different methods were used for the synthesis of colloidal W-ZnO nanoparticles doped
with TM ions such as Co2+ or Mn2+ , and W-CoO nanocrystals.
Pure ZnO and Zn1−x TMx O with x <30% for Co and x <2% for Mn QDs were prepared
using hydrolysis and condensation of acetates ((OAc)2 ) solution in dimethyl sulfoxide
(DMSO) [29], [30]. The experimental procedure was as follows: Acetates were dissolved
in DMSO, typically 9·10−3 moles of Zn(OAc)2 and Co(OAc)2 or Mn(OAc)2 in 90 ml
of DMSO. Simultaneously, tetramethyl ammonium hydroxide (N(Me)4 OH) (1.8 equiv
of OH− ) was dissolved in 30 ml of ethanol. Then, the N(Me)4 OH solution was added
dropwise to the (OAc)2 in DMSO under constant stirring and the reaction took place.
After a rapid nucleation, the reaction Ostwald ripens, which means that small crystals
are unstable and shrink due to their large surface-to-volume ratio while larger particles
are stable and grow [31]. The reaction time and the addition of N(Me)4 OH are important
factors in this synthesis. Bigger nanoparticles were obtained for longer reaction times.
When Co(OAc)2 or Mn(OAc)2 concentration increases, more time is needed, not only for
the reaction to occur, but also for the nanoparticles to grow. Dopants are excluded from
the initial nucleation but are incorporated into the ZnO nanoparticles almost isotropically
during growth from solution. The formation of ZnO becomes linear with base addition
above 1.0 equiv of OH− . Figure 3.4. shows the sequence of a reaction flask during the
synthesis of a Co2+ -doped ZnO sample. The first image, before the base addition, shows
the characteristic pink color of octahedral Co2+ in (OAc)2 . N(Me)4 OH addition turns
the solution from pink to blue, indicating the tetrahedral coordination of Co2+ in ZnO.
Nanocrystals prepared by this method were precipitated in ethyl acetate, washed with
ethanol and precipitated again by adding heptane. Colloidal nanocrystals were usually
stabilized by a layer of surfactants attached to their surface. The energy with which
surfactant molecules adhere to the surfaces of growing nanocrystals is one of the most
important parameters influencing crystal growth. Murray et al. discussed this fact for
46 3.7. Colloidal semiconductor nanocrystals
the growth of CdSe nanoparticles capped with trioctylphosphine oxide (TOPO) [32]. For
the final capping treatment, melted TOPO was added to the nanocrystals and heated at
180 ◦ C for 30 minutes. The resulting powders were precipitated in ethanol or methanol
and resuspended in toluene.
Figure 3.4: Preparation of ZnO: Co2+ colloidal nanocrystals at the Chemistry Department, University
of Washington. The solution changes from pink to blue indicating the conversion of octahedral Co2+ to
tetrahedral Co2+ .
Colloidal W-CoO nanoparticles were synthesized using a nonaqueous solution method

with the reaction system composed by the metal fatty acid salt, the corresponding fatty
acid and a hydrocarbon solvent [33], [30]. For preparing CoO nanoparticles, cobalt
stearate (Co(St)2 ) (0.625 g) was used as the precursor, a certain amount of stearic acid
(0.305 g) was employed as ligands and octadecene (ODE) (5 g) was chosen as the non-
coordinating solvent. Co(St)2 and stearic acid were dissolved in ODE stirring at 120 ◦ C
under vacuum. When the purple mixture was fully dissolved, it was placed in a N2 flow,
and it was rapidly heated to 320 ◦ C and held there under stirring for 1 hour. When the
solution turned greenish W-CoO nanocrystals were formed. After cooling to 100 ◦ C, the
obtained CoO particles were precipitated with ethanol, capped with TOPO, precipitated
in acetone and suspended in toluene.
Bibliography
[1] M.P. Pechini. U.S.A. Patent 3.330.697. 1967.
[2] H. Flachsbart and W. Stöber. Preparation of Radioactively Labeled Monodisperse

Silica Spheres of Colloidal Size. J. Colloid Interface Sci., 30: 568–573, 1969.

[4] S.R. Mishra, G.J. Long, F. Grandjean, R.P. Hermann, S. Roy, N. Ali, and A. Viano.
Magnetic properties of iron nitride-alumina nanocomposite materials prepared by
high-energy ball milling. Eur. Phys. J. D, 24: 93–96, 2003.
[5] V.V. Boldyrev. Mechanochemistry and mechanical activation of solids. Solid State
Ionics, 63-65: 537–543, 1993.
[6] S. Mørup, J.Z. Jiang, F. Bødker, and A. Horsewell. Crystal growth and the steady-
state grain size during high-energy ball-milling. Europhys. Lett., 56: 441–446, 2001.
[7] R.E. Thoma, G.M. Hebert, H. Insley, and C.F. Weaver. Phase equilibria in the
System Sodium Fluoride-Yttrium Fluoride. Inorg. Chem., 2: 1005–1012, 1963.
[8] N. Martin, P. Boutinaud, R. Mahiou, J.C. Cousseins, and M. Bouderbala. Prepa-

ration of fluorides at 80 ◦ C in the NaF(Y, Yb, Pr)F3 system. J. Mater. Chem,
9: 125–128, 1999.
[9] J. Lu, Q. Zhang, and F. Saito. Mechanochemical Synthesis of Nano-sized Complex

Fluorides from Pair of Different Constituent Fluoride Compounds. Chem. Lett.,
1: 1176–1177, 2002.
47
48 BIBLIOGRAPHY
[10] T. Tsuzuki and P.G. McCormick. Mechanochemical synthesis of metal sulphide

nanoparticles. Nanostruct. Mater., 12: 75–78, 1999.
[11] T. Tsuzuki and P.G. McCormick. Synthesis of CdS quantum dots by mechanochem-
ical reaction. Appl. Phys. A, 65: 607–609, 1997.
[12] T. Tsuzuki, J. Ding, and P.G. McCormick. Mechanochemical synthesis of ultrafine

zinc sulfide particles. Physica B, 239: 378–387, 1997.
Er3+ . J. Phys. Chem. B, 106: 1181–1187, 2002.
[14] J.A. Capobianco, F. Vetrone, T. D’Alessio, G. Tessari, A. Speghini, and M. Bettinelli.

Optical spectroscopy of nanocrystalline cubic Y2 O3 : Er obtained by combustion
synthesis. Phys. Chem. Chem. Phys., 2: 3203–3207, 2000.
[15] T. Ye, Z. Guiwen, Z. Weiping, and X. Shangda. Combustion synthesis and photolu-
minescence of nanocrystalline Y2 O3 : Er phosphors. Mater. Res. Bull., 32: 501–506,
1997.

[17] A.V. Rosario and E.C. Pereira. The effect of composition variables on precursor
degradation and their consequence on Nb2 O5 film properties prepared by the Pechini
Method. J. Sol-Gel Sci. Techn., 38: 233–240, 2006.
[18] J.S. Steckel, J.P. Zimmer, S. Coe-Sullivan, N.E. Stott, V. Bulovic, and M.G.
Bawendi. Blue Luminescence from (CdS)ZnS Core-Shell Nanocrystals. Angew.
Chem., 116: 2206–2210, 2004.
[19] O. Lehmann, K. Kömpe, and M. Haase. Synthesis of Eu3+ -Doped Core and
Core/Shell Nanoparticles and Direct Spectroscopic Identification of Dopant Sites at
BIBLIOGRAPHY 49
the Surface and in the Interior of the Particles. J. Am. Chem. Soc., 126: 14935–14942,
2004.
[20] D. Hreniak, P. Psuja, W. Strȩk, and J. Hölsä. Luminescence properties of Y3 Al5 O12 :
Eu3+ -coated submicron SiO2 particles. J. Non-Cryst. Solids, 354: 445–450, 2008.
[21] P. Zhu, Q. Zhu, H. Zhu, H. Zhao, B. Chen, Y. Zhang, X. Wang, and W. Di. Effect
of SiO2 coating on photoluminescence and thermal stability of BaMgAl10 O17 : Eu2+
under VUV and UV excitation. Opt. Mat., 30: 930–934, 2008.
[22] W. Stöber, A. Fink, and E. Bohn. Controlled Growth of Monodisperse Silica Spheres
in the Micron Size Range. J. Colloid Interface Sci., 26: 62–69, 1968.
[23] Q. Lü, A. Li, F. Guo, L. Sun, and L. Zhao. The two-photon excitation of SiO2 -
coated Y2 O3 : Eu3+ nanoparticles by a near-infrared femtosecond laser. Nanotech.,
19: 205704–205712, 2008.
[24] W. Ge, H. Zhang, J. Wang, D. Ran, S. Sun, H. Xia, J. Liu, X. Xu, X. Hu, and
M. Jiang. Growth and thermal properties of Co2+ : LaMgAl11 O19 crystal. J. Crystal
Growth, 282: 320–329, 2005.
[25] P.Y. Jia, M. Yu, and J. Lin. Sol−gel deposition and luminescent properties of
LaMgAl11 O19 : Ce3+ /Tb3+ phosphor films. J. Solid State Chem., 178: 2734–2740,
2005.
[26] S. Ekambaram, K.C. Patil, and M. Maaza. Synthesis of lamp phosphors: facile
combustion approach. J. Alloys Compd., 393: 81–92, 2005.
[27] J.M.P.J. Verstegen, J.L. Sommerdijk, and J.G. Verriet. Cerium and Terbium lumi-
nescence in, LaMgAl11 O19 . J. Lumin., 6: 425–431, 1973.
2006.
50 BIBLIOGRAPHY
[31] Y. Yin and A.P. Alivisatos. Colloidal nanocrystal synthesis and the organic-inorganic
interface. Nature, 437: 664–670, 2005.
[32] C.B. Murray, D.J. Norris, and M.G. Bawendi. J. Am. Chem. Soc., 115: 8706–8715,
1993.
[33] N.R. Jana, Y. Chen, and X. Peng. Size- and Shape-Controlled Magnetic (Cr, Mn,
Fe, Co, Ni) Oxide Nanocrystals via a Simple and General Approach. Chem. Mater.,
16: 3931–3935, 2004.
Chapter 4
Experimental Methods
4.1 Introduction
Along this chapter, we show the experimental procedures and equipment utilized not
only for the structural characterization of the samples, but also for the spectroscopic
study developed. First, XRD and transmission electron microscopy (TEM) techniques
are detailed. Then, the different setups used for the spectroscopic measurements are
described. Absorption, luminescence, excitation, lifetime and Raman measurements have
been performed. Besides, low temperature and high pressure techniques are described.
Finally, the spectra correction procedures are explained.
4.2 Structural characterization
Different techniques such as XRD or TEM were employed to characterize the prepared
materials. XRD measurements were accomplished in order to check the phase purity of all
samples and to estimate the particle size in the case of nanocrystalline samples. Average
particle size and size distribution of the nanoparticles were also estimated from TEM
images.
4.2.1 X-ray diffraction
XRD technique is commonly applied for the structural characterization of all kind of
materials. It is based on the diffraction produced when a X-ray beam is scattered by a
material. If the material presents crystalline structure, the scattered waves interfere in a
51
constructive way for specific directions according to Bragg’s law, 2d sin θ = nλ, where d
is the distance between crystallographic planes, λ is the incident X-ray wavelength, θ is
the angle between the incident wave and the dispersion plane, and n is an integer number
indicating the diffraction order.
Various X-ray diffractometers were used to obtain XRD patterns (see Fig. 4.1). A
Philips 1700 diffractometer with a Bragg-Brentano θ − 2θ geometry, in which the X-ray
source is fixed, the sample turns θ and the detector turns 2θ, was utilized at the Solid State
Physics laboratory (University of Cantabria). A Thermo Electron ARL X’TRA based on
a vertical θ − θ geometry, where both the X-ray source and the detector move an angle
θ, was employed during a collaboration with Prof. Marco Bettinelli at the Biotecnology
Department (University of Verona). Both diffractometers are equipped with a Cu-anode
X-ray source (Kα , λ=1.5418 Å). The phase of the samples synthesized at Verona was
identified by comparing the obtained XRD patterns with the PDF 4+ 2006 database and
using SIeve program, while MAUD program was used for Rietveld data refinement [1].
Figure 4.1: X-Ray diffractometers employed to characterize our samples, Philips 1700 (left) and Thermo
Electron ARL X’TRA (right).
High pressure XRD experiments were developed with a Xcalibur diffractometer (Ox-
ford Diffraction Limited) (Fig. 4.2) at the Complutense University of Madrid within the
MALTA-Consolider project. XRD patterns were obtained on a 135 mm Atlas charge-
coupled device (CCD) detector placed at a distance of 90 mm from the sample using
Kα1 : Kα2 molybdenum radiation (λ=0.7107 Å). The X-ray beam was collimated to a
diameter of 300 µm. Exposure times were generally of 1 hour and the accessible angular
range is 4θ=60◦ . The observed intensities were integrated as a function of 2θ in order
to give conventional one-dimensional diffraction profiles. The CrysAlis software (Oxford
Diffraction Limited) was used for the data collection and the preliminary reduction of
the data. The indexing and refinement of the powder patterns were performed using the
TOPAS package.
Figure 4.2: Xcalibur diffractometer (Oxford Diffraction Limited) used for the high pressure measure-
ments (left). A CCD image of a CdS XRD pattern under pressure is also shown (right).
The particle size can be determined by X-ray peak broadening analysis for nanocrys-
talline samples with size up to 500 nm [2]. The peaks diffraction broadening arises mainly
due to three factors: Instrumental effects, crystalline size and lattice strains. The peak
width due to instrumental broadening is estimated using a standard Si sample. If the
observed XRD peak has a width B0 , and the width due to instrumental effects is Bi ,
then Br , which corresponds to XRD peak width at half-maximum once the instrumental
q p
broadening has been substracted, Br = (B0 − Bi ) B02 − Bi2 , is due to both crystalline
size and lattice strains. Broadening of XRD peaks due only to crystalline size is given by
kλ
Scherrer formula, Bc = L cos θ
, where λ is the X-ray wavelength, θ is the Bragg angle, L
is the average particle size and k is a constant. The broadening caused by lattice strains
can be represented by Bs = η tan θ, where η is the strain in the material. From these
expressions, the Williamson-Hall equation to determine the particle size is deduced [3]:
kλ kλ
Br = Bc + Bs = + η tan θ → Br cos θ = + η sin θ (4.1)
L cos θ L
When Br cos θ is plotted against sin θ a straight line is obtained; crystalline size can
kλ
be calculated from the intercept value L
[2]. The Williamson-Hall approach separates
the effect of size and strain in the nanocrystals. Spherical shape is assumed for the
nanoparticles. Strain contribution to the line broadening is significant for nanoparticles
grown by mechanical procedures.
4.2.2 Transmission electron microscopy
TEM is a technique of great importance as far as structural characterization of nanometric

systems is concerned. In TEM measurements, electrons are accelerated and focused on the
specimen (sample) by the condenser system, usually two lenses. After passing through the
sample, electrons are collected and the objective lenses form both images and diffraction
patterns. The main parts of the microscope are: Electron source, sample holder, light
and electron optics, electron detection and display. The TEM resolution, the smallest
0.61λ
distance that can be resolved δ, is given by δ = β
, where λ is the wavelength of the
radiation and depends on the electrons energy, and β is the semi-angle of collection of the
magnifying lenses. One important drawback of the TEM is that only a small part of the
sample is seen at a time. Moreover, TEM presents 2D images of 3D specimens, so the
information is averaged through its thickness. Another limitation is related to the fact
that samples must be ”electron transparent”, which means they must be thinner than
100 nm. For our powder specimen, preparation is easier than for thin films or metallic
samples, and involves selecting fine powder, suspending it in a non reactive liquid, and
placing it on the grid [4].
TEM experiments were performed on a JEM 2100 electron microscope (JEOL, Japan)
(Fig. 4.3) at the TEM facility at the University of Cantabria (SERMET), operating at
an accelerating voltage of 200 kV, with an estimated point to point resolution of 0.23
nm, and equipped with a CCD detector (Gatan Orius SC 1000B ). DigitalMicrograph
(Gatan) software was employed for the data collection and treatment. The samples were
prepared by suspending the solid powder in ethanol (Panreac, 96 purity) under ultrasonic
vibration. One drop of the prepared suspension was applied to carbon films on copper
grids (Agar Scientific).
Figure 4.3: JEM 2100 transmission electron microscope (JEOL, Japan) used at the TEM facility at
the University of Cantabria.
4.3 Spectroscopy
The aim of this work is the study of the optical properties of the synthesized materials.
Different spectroscopic techniques were used to analyze the involved transitions in the
light-matter interaction. As it is described below, absorption, luminescence, excitation,
time resolved spectroscopy, lifetime and Raman spectroscopy techniques were applied.
The spectroscopic measurements were carried out at the High Pressure and Spectroscopy
group laboratories at the University of Cantabria.
4.3.1 Absorption
A light beam becomes attenuated in a material according to Lambert-Beer law, I =

I0 exp(−αx), where I0 is the incoming intensity, I is the light intensity after passing
through a thickness x of the sample, and α is the absorption coefficient. Absorption
spectra involve the measurement of the light attenuated by a sample as a function of
the radiation wavelength, A(λ). The absorbance is defined as A = log II0 . In the case of
powder materials, absorption can only be measured by using very thin samples, although
reflectance spectra provide similar information. The spectral reflectance is defined as the
ratio of the flux reflected by the specimen to that of a standard surface under identical
geometrical and spectral conditions. Reflectance spectra can be registered in two different
modes: specular reflectance and diffuse reflectance, for the latter, an integrating sphere
is required to collect the diffuse reflected light for different wavelengths [5].
Figure 4.4: Cary 6000i spectrophotometer employed for reflectance measurements. The optical design
of the diffuse reflectance sphere (a) and the two components of the reflected light, specular and diffuse
reflection (b) are shown.
A Cary 6000i (Varian) spectrophotometer was used to register diffuse reflectance spec-
tra in the range 200-1800 nm (Fig. 4.4). It is equipped with two light sources; a quartz
halogen lamp for the visible/IR region and a deuterium lamp for the UV, and two de-
tectors; one photomultiplier (PMT) (Hamamatsu R928 ) for the visible region and one
InGaAs detector for the near IR. Diffuse reflectance measurements were performed using a
polytetrafluoroethylene (PTFE)-coated integrating sphere. Initially, a baseline is recorded
with the PTFE reference disk covering the reflectance port. The sample is then mounted
over the port and the light reflected by the sample surface is collected by the sphere.
The total (diffuse and specular) or the diffuse-only reflectance (see Fig. 4.4(b)) may be
measured by mounting the sample against the sphere port in two different configurations.
A FT-IR System 2000 (Perkin-Elmer) spectrophotometer was used to obtain the trans-
mission spectra in the IR region. It consists essentially of three components: Two radi-
ation sources to cover the 4400-400 cm−1 range, one Michelson interferometer and a
Mercury-Cadmium-Telluric detector. In addition, IR absorption spectra of the samples
in the form of KBr pellets were recorded on another spectrophotometer (Nicolet Magna
760 ), with a Deuterated-Triglycine-Sulfate detector.

Absorption experiments at high pressure requires non-conventional home-made se-
tups. The experimental apparatus used in this work is shown in Fig. 4.5. Tungsten
and deuterium lamps attached to a monochromator (Acton Research Corporation Spec-
tra Pro-300i) were used to obtain monochromatic light. One parabolic mirror and two
reflection objectives were incorporated in order to avoid chromatic aberration. The first
objective focuses the excitation beam in the hydrostatic cavity and the second one collects
the transmitted light, which is measured using a PMT with a lock-in technique (Stanford
Research systems Sr830 DSP ). A spatial filter must be used before the first objective.
Figure 4.5: Optical setup to measure high pressure absorption at the High Pressure and Spectroscopy
group laboratories, University of Cantabria.
4.3.2 Luminescence and excitation
In luminescence spectra, the emitted light intensity is measured as a function of the wave-
length for a fixed excitation wavelength, I(λ). On the contrary, in excitation technique,
only one wavelength is detected and the excitation wavelength is scanned in a certain
spectral range. Both luminescence and excitation, in contrast to absorption, are selective
techniques. Different light sources in the IR-visible-UV range, diverse detection systems
sensitive to those spectral regions and several monochromators were used to obtain this
kind of spectra. All spectra were corrected for the system response and were represented
as photons counts vs wavenumbers (see Section 4.6).
A standard spectrofluorimeter (Jobin-Yvon Fluorolog-2 ) was employed for the mea-

surement of photoluminescence and excitation spectra at room temperature (RT) and low
temperatures (Fig. 4.6). The sample is excited with a Xe-lamp followed by an excita-
tion monochromator, and the emitted light is detected with a PMT (Hamamatsu R928 )
at a 90◦ configuration in order to avoid direct light from the lamp. Continuous wave
(CW) UC luminescence spectra were obtained by exciting with a CW laser-diode (LD)
(LUMICS GmbH & Arroyo Instruments) and using the Fluorolog-2 detection system (see
setup scheme in Fig 4.6(a)). The LD can reach a maximum output power of 4 W and its
emission wavelength can be slightly tuned by changing the temperature. To obtain the
spectra, the nanopowders were transferred into a quartz capillary and closed after partial
air evacuation. A special sample holder was designed and adapted to the Fluorolog-2 (Fig.
4.6(b)). This allows to measure all the samples in identical conditions. Yb3+ luminescence
spectra were recorded by exciting with the LD in a single monochromator equipped with
an extended IR PMT (Hamamatsu R7102 ).
Figure 4.6: Jobin-Yvon Fluorolog-2 fluorimeter used for luminescence and excitation measurements.
The adapted setup for temperature-dependent UC experiments upon excitation with a LD (a) and the
sample holder (b) are also shown.
Another setup utilized to obtain luminescence spectra is shown in Fig. 4.7. An optical
parametric oscillator (OPO) laser system (Opotek Vibrant model B 355 II ) pumped by
the third harmonic of a Q-switched Nd:YAG (Brilliant Quantel) was used as excitation
source. This tunable OPO system offers an approximate pulse width of 10 ns with a
repetition rate of 10 Hz, and a continuous tuning range of 410-710 nm in the visible region
(signal mode) and 710-2400 in the IR region (idler mode). Pulses were guided through
a coupled optic fiber to the sample. The emitted light is collected by a 10× Mitutoyo
objective and guided through a fiber to the detection system. A monochromator (TRIAX
320) and an intensified CCD camera (Horiba-Jobin-Yvon iCCD 3553 ) were used for the
visible and UC luminescence intensity detection.
Figure 4.7: Optical setup used for visible and UC luminescence experiments at the High Pressure and
Spectroscopy group laboratories, University of Cantabria.
4.3.3 Lifetime and time resolved spectroscopy
To obtain information about both radiative and non-radiative decay from an excited state
and energy transfer processes, the luminescence temporal evolution after pulsed excitation
must be measured. This was performed in two different manners:
• Recording the temporal dependence of the luminescence intensity at a specific wave-

length. For the simplest case, in which excitation and emission occur in the same
center, an exponential decay is measured. Fitting this curve to a single exponen-
tial, I(t) = I0 exp( −t
τ
), the excited state lifetime, τ , is obtained. On the contrary,
when energy transfer processes are involved, an intensity rise followed by a decay is
detected once the excitation pulse has stopped.
• Obtaining the whole emission spectra at different times after the excitation pulse
has reached the sample. This experimental procedure is called time-resolved lumi-
nescence and the emission spectrum is recorded at a certain delay time with respect
to the excitation pulse and within a temporal gate. This technique is very useful to
explore energy transfer processes and allows us to identify contributions of differ-

ent optically active centers present in a sample if they have different desexcitation
dynamics.
Figure 4.8: Optical setup for lifetime measurements at the High Pressure and Spectroscopy group
For fluorescence lifetime experiments, the 10 ns laser pulses of the OPO system or
modulated LD excitation were used. In the case of this last device, the modulation
frequency must be suitable so that the system can reach the steady state and the lifetime
can be measured. The setup employed for lifetime measurements upon OPO pulsed
excitation is shown in Fig. 4.8. The sample luminescence was dispersed by a 0.50 m single
monochromator (CHROMEX 500IS/SM) equipped with 500 nm blazed 1200 grooves/mm
and 750 nm blazed 600 grooves/mm gratings, detected by a PMT (Hamamatsu R928 ) or
extended IR PMT (Hamamatsu R7102 ) and recorded with a multichannel scaler (Stanford
Research SR-430). An appropriate load resistance was placed between both devices. The
time constant of the RC circuit (C = 10 pF ) should be much lower than the luminescence
lifetime, τ , in order to avoid deconvolution processes. The smallest load resistance, R =
20 Ω was used in all measurements. The OPO+iCCD setup described in the previous
section (fig. 4.7) was also used for lifetime and time-resolved luminescence experiments.
4.3.4 Raman
Raman spectroscopy deals with the study of the fraction of incident light scattered by a
sample. Rayleigh scattering is an elastic photon process in which the scattered photon
energy is equal to the incident photon energy. However, in Raman spectroscopy the light
is inelastically scattered by a substance. The Stokes and anti-Stokes Raman scattering
involve virtual levels which do not correspond to real states, as a result, Raman spectra are
much weaker than fluorescence spectra by a factor of about 106 −108 . Raman experiments
are usually carried out under non-resonant illumination so that the Raman spectrum is
not masked by the more efficient emission spectrum [5], [6]. A typical Raman spectrum
shows an intense band at the incident frequency, ω0 , resulting from Rayleigh scattering,
and fainter Raman bands on both sides of ω0 at distances corresponding to vibrational
frequencies, ωk . In practice, the Stokes Raman scattering is more intense than the anti-
Stokes and the lower spectral region is used [7].
Figure 4.9: Equipment used for Raman measurements at the High Pressure and Spectroscopy group
The T64000 Raman spectrometer system (Horiba), together with a Kripton-Argon

laser (Coherent Innova Spectrum 70C) (the 514.5 nm green line was used), and a Nitro-
gen cooled CCD (Jobin-Yvon Symphony) with a confocal microscopy for the detection
(Fig. 4.9), were employed for Raman experiments. The T64000 system is composed of
three monochromators (640 mm focal length) with two basic configurations; triple addi-
tive mode, with the three monochromators operating in series (3×640 mm focal length),
62 4.4. Temperature dependence
and substractive mode, in which the two first monochromators filter the laser, and the
analysis is done with the third one. Different high pressure cells can be adapted to the
XYZ automated stage for high pressure measurements. LabSpec software permits data
acquisition and treatment.
4.4 Temperature dependence
Many of the spectroscopic characterization measurements, such as luminescence, excita-

tion or lifetime, were accomplished as a function of temperature.
• Low temperature measurements were achieved using a closed-cycle helium cryo-

stat (Air Products CS202E ) (Fig. 4.10). The major components are the expander
or cold finger, the compressor, the radiation shield and the vacuum shroud. The
sample is placed on the cold finger using Cry-Con copper-filled thermal grease to
fix the sample and to ensure thermal conductivity at low temperatures. This sys-
tem requires a vacuum pump for the sample space. A combination of a diffusion
pump (Leybold PD 180 L) with a rotatory pump (Leybold Trivac B ) as well as a
turbomolecular pump (Varian Turbo Dry 70 ) were employed in our experiments.
A programmable temperature controller (APD-K cryogenics HC-2 ) together with
a Si-diode thermocouple let us modify and stabilize the temperature in the range
10-300 K with an accuracy better than 0.1 K.
Figure 4.10: Closed-cycle helium cryostat components (left) and compressor (right) used for low tem-
perature experiments down to 15 K.
• High temperature experiments (300-650 K) were performed using a microscope heat-

ing stage (Leitz 350) (Fig. 4.11). The sample is placed between two quartz films and
put into a metallic cavity. The heat is controlled by a rheostat and a thermocouple
is used to measure the temperature.
Figure 4.11: Heating stage, power supply and thermometer for the temperature measurements used in
the range 300-650 K.
4.5 High pressure measurements

High pressure studies are a valuable mechanism to explore changes in materials as a
function of volume without changing the chemical composition. Hydrostatic pressure
is used to systematically influence the bonding environment of luminescent centers and
therefore modify electronic states energy and transitions probabilities. Sometimes, high
pressure can induce structural phase transitions and even stabilize structures not achiev-
able through other means. As a consequence, their optical properties change [8]. High
pressure is also a very efficient tool for understanding the electronic structure of semi-
conductors. The interatomic distance is reduced in average when hydrostatic pressure
is applied to a material, and this fact normally causes an increase in the semiconductor
bandgap energy [9].
4.5.1 Diamond anvil cells
High pressure measurements were carried out in diamond anvil cells (DACs). The use of
diamond has two important advantages over other anvils. Firstly, diamond is the hardest
known material and is capable of reaching higher pressures [10]. Secondly, diamond is
transparent in a wide frequency range including X-ray and IR-visible range.
Figure 4.12: Hydrostatic cavity in the cell obtained after placing the perfored gasket containing the
sample between the two opposite diamond anvils.
The basic principle of the DAC is very simple. A metal gasket containing the sample,
the hydrostatic medium and ruby chips is placed between the flat parallel faces of two
diamonds (see Fig. 4.12). By pushing the two opposed anvils together, the sample pressure
increases. The gasket is prepared by drilling a hole at the center of the indentation made
by the anvil face. Then, the sample, the medium, and a ruby chip if necessary, are placed
inside the cavity [11].
Figure 4.13: Spark eroder used for gaskets perforation. This system allows the utilization of wires
between 25 and 400 µm. A picture of a perfored gasket is also shown.
Sample chamber preparation.
The metal foil or gasket serves three purposes; it provides the high-pressure sample cham-
ber, it avoids direct contact between diamonds, and it gives lateral support to the conical
faces of the anvils. The gasket is first indented by an anvil face, this process compresses
and hardens the material. The initial thickness of the foil is about 300 µm whereas the
indented thickness varies from 50 to 200 µm. The deeper the indentation is, the higher
pressures can be reached. Gasket perforation was performed using a semi-automatic spark
erosion machine (Betsa MH 20M) with tungsten electrodes (Fig. 4.13). The hydrostatic
cavity diameter was usually between 150 and 300 µm.
DACs specifications.
Different cells were used in the high pressure experiments at the High Pressure and Spec-
troscopy group laboratories at the University of Cantabria, as well as at the MALTA
X-Ray Diffractometer at the Complutense University of Madrid. A Brillouin-Raman cell
(Diamond Optics Inc.) (Fig. 4.14) and a Membrane cell (developed at the Pierre and
Marie Curie University) (Fig. 4.15) were utilized for absorption and Raman measurements
on CdS and ZnO nanocrystals under high pressure. A cryoDAC-Mega cell (easyLab Tech-
nologies) was also employed for high pressure luminescence and lifetime experiments on
GGG: Cr3+ nanoparticles. High pressure XRD measurements on CdS nanocrystals were
performed in a modified Merrill-Bassett (MALTA cell).
• Brillouin-Raman cell This cell is based on the modification made by Mao and
Bell of the Merril-Basset cell, and is designed for reaching pressures as high as 40
GPa [11], [12], [13]. The main body is made of two 440c steel pieces perfectly face to
face in which the two plates are fit. Diamond anvils are placed in the middle of the
plates. Pressure is applied by turning four allen screws which pull the two plates,
and hence the diamonds, together. Type-IIa diamonds with very low luminescent
impurities concentration, and with a 0.6 mm culet diameter, were used for the anvil.
• Membrane cell The main difference between the membrane cell and the Brillouin-
Raman cell is that the former uses a metallic diaphragm or membrane to generate
the pressure instead of screws. This cell is composed of two steel parts which fit
perfectly one into the other. The membrane, a toroidal cavity in which pressurized
gas is introduced, is mounted on the upper part of the cell. When the membrane
is inflated, it pushes the upper cylinder against the other increasing the pressure
Figure 4.14: Diamond Optics Brillouin-Raman cell used for absorption and Raman measurements at
high pressures.
Figure 4.15: Membrane cell used for absorption and Raman measurements at high pressures.
in the hydrostatic cavity. This system allows a fine pressure control and has the
advantage that there is no need of moving the cell to change the pressure. The
utilized diamond anvils had 0.5 mm of culet size.
• cryoDAC-Mega cell This cell, made of CuBe alloy, is suitable for both optical and
XRD studies at cryogenic temperatures. The diamond anvils are mounted within
CuBe rings and mechanically fixed to their tungsten carbide support plates. The
DAC itself is held within two CuBe clamps, and one of them is fixed to the cryostat.
Pressure is applied by turning the four bolts on the CuBe blocks.
• MALTA cell. In the case of this modified Merrill-Bassett DAC, three screws pull
the two plates together [11]. The used diamond anvils had 1 mm of culet size.
4.5.2 Pressure calibration and transmitting media
The most commonly used technique to determine the pressure inside the DAC cavity is
based on the pressure shift of the R-lines (2 E → 4 A2 ) ruby luminescence. The RT R1 and
R2 shift rate is linear up to 20 GPa according to [14]:
ER1 (cm−1 ) = 14405 − 7.53 · P (GPa)
ER2 (cm−1 ) = 14434 − 7.53 · P (GPa) (4.2)
Ruby shows a sharp and efficient luminescence facilitating its detection. However,
there are times, like for instance in the case of GGG: Cr3+ , when ruby emission can hide
the sample luminescence, and a non-luminescent pressure sensor must be used. Spectral
shift of Raman lines in diamond have also been calibrated as a function of pressure
[15]. The Raman-active phonon energy is DR = 1332.26 cm−1 at ambient pressure (AP).
Pressure can be expressed as a function of the Raman peak position [16]:
DR (cm−1 ) = 1332.26 + 2.64 · P (GPa) (4.3)
The function of the transmitting medium is to ensure a homogeneous pressure dis-

tribution in the sample chamber. Pressure gradients and stress must be reduced, since
they can alter the physical state of the sample independently of any hydrostatic pressure
effects. Basic requirements for the hydrostatic medium are being transparent in the wave-
length range under study, and keeping the hydrostaticity within a wide pressure interval.
Different transmitting media were used. Paraffin oil is hydrostatic below 10 GPa and it
was chosen as a good candidate because it is inert. When higher pressures were required,
silicon oil (Dow Corning) was employed. A 4:1 mixture of ethanol: methanol provides
hydrostatic conditions at RT up to 10 GPa and it was also used.
4.6 Spectra correction

The correction of the experimental excitation and luminescence spectra is important when
precise information about spectral positions, bands shape and intensity is desired. Two
types of corrections were applied in this work.
4.6.1 Wavelength correction
In the spectra shown in this work, the x axis is represented as wavenumbers, in cm−1
units. However, the raw data were obtained as a function of wavelength, λair (nm). To
accomplish this conversion, first, we need to calibrate the monochromator using a Hg
lamp. Then, the refractive index of air must be considered in order to get the vacuum
wavelength, λvac (nm), according to the following expression [17]:
1.2288 3.555 · 104

λvac = λair + 2.72643 · 10−4 · λair + + (4.4)
λair λ3air
Finally, the vacuum wavelength, λvac (nm), is converted into wavenumbers, E(cm−1 ):
E(cm−1 ) = 107 · λ−1

vac (nm) (4.5)
4.6.2 Intensity correction
• Excitation spectra. In order to correct the excitation spectra intensity, we must

bear in mind the different radiation source power as a function of the wavelength,
the response R(λ). The excitation spectra in the visible region were achieved using
a Xe-lamp in the Fluorolog-2. The lamp power is measured simultaneously while a
spectrum is being collected. Yb3+ excitation spectra were recorded exciting with the
OPO laser. The laser power is monitored using a beam splitter and a powermeter
(Ophir 3A-SH ). In all cases, the excitation spectra were corrected by dividing them
by the corresponding response.
• Emission spectra. When luminescence spectra are measured, the detector signal
is determined by the amount of light emitted by the sample at each wavelength, and
the efficiency with which that light is detected. The experimental spectra must be
corrected so that they can be interpreted only in terms of the intrinsic emission of
the sample. The experimental procedure described by Ejder was followed [18]. The
signal can be expressed in terms of I (energy per second) or J (photons per second)
for y axis, and either λ(nm) or E(cm−1 ) for x axis. In this work, all the spectra are
represented as photons vs wavenumbers, so the integrated intensity of an emission

band is proportional to the number of emitted photons per time unit. First, the
spectrum of a black body source (calibrated filament lamp) was measured at the
same experimental conditions, B(λ). Then, the lamp spectrum was divided by the
theoretical output of the lamp, J(E):
1
J(E) ∝ E 2 (4.6)
exp( kBET ) −1
where T is the black body temperature (2535 K), and the detection system response
was obtained, R:
B(E)
R= (4.7)
J(E)
To correct a luminescence spectrum initially expressed in J(λ), LS (nm), one must

multiply by λ2 (nm), LS ’(nm)=LS (nm)·λ2 (nm), convert x axis to wavenumbers,
LS (cm−1 ) and divide by the response curve, R. The corrected emission spectrum is
given by LSc :
LS(cm−1 )
LSc = (4.8)
R
Bibliography
[1] L. Luterotti and S. Gialanella. X-ray diffraction characterization of heavily deformed

metallic specimens. Acta. Mater., 46: 101–110, 1998.
[2] C. Suryanarayana and M.G. Norton, editors. X-Ray Diffraction: A Practical Ap-
proach. Plenum Press, New York, 1998.
[3] G.K Williamson and W.H Hall. X-ray line broadening from filed aluminium and
wolfram. Acta. Metall., 1: 22–31, 1953.
[4] D.B. Williams and C.B. Carter, editors. Transmission Electron Microscopy. Plenum
Press, New York, 1996.
[9] T. Suski and W. Paul, editors. Semiconductors and Semimetals, volume 54 of High
Pressure in Semiconductor Physics I. Academic Press, San Diego, 1998.
71
72 BIBLIOGRAPHY
[10] I.V. Aleksandrov, A.F. Goncharov, A.N. Zisman, and S.M. Stishov. Diamond at high
pressures: Raman scattering of light, equation of state, and high pressure scale. Sov.
Phys. JETP, 66: 384–390, 1987.
[11] A. Jayaraman. Diamond anvil cell and high-pressure physical investigations. Reviews
of Modern Physics, 55: 65–108, 1983.
[12] W.B. Holzapfel and N.S. Isaacs, editors. High-pressure techniques in Chemistry and
Physics. Oxford University Press, New York, 1997.
[13] W.F. Sherman and A.A. Stadtmuller, editors. Experimental techniques in high-
pressure research. John Wiley and Sons, New York, 1987.
[14] G.J. Piermarini, S. Block, J.D. Barnett, and R.A. Forman. Calibration of the pressure
dependence of the R1 ruby fluorescence line to 195 kbar. J. Appl. Phys., 46: 2774–
2780, 1975.
[15] D. Schiferl, M. Nicol, J.M. Zaug, S.K. Sharma, T.F. Cooney, S.Y. Wang, T.R. An-
13 12
thony, and J.F. Fleischer. The diamond C / C isotope raman pressure sensor
system for high-temperature/pressure diamond-anvil cells with reactive samples. J.
Appl. Phys., 82: 3256–3265, 1997.
[16] A. Tardieu, F. Cansell, and J.P. Petitet. Pressure and temperature dependence of
the first-order Raman mode of diamond. J. Appl. Phys., 68: 3243–3245, 1990.
[17] D.R. Lide, editor. Handbook of Chemistry and Physics. 65th edition. CRC Press,
1983.
[18] E. Ejder. Methods of Representing Emission, Excitation and Photoconductivity

Spectra. J. Opt. Soc. Am., 59: 223–224, 1969.
Chapter 5
Insulating materials. Luminescent

properties
5.1 Introduction
In this chapter, we present the most relevant results in relation with the electronic tran-
sitions responsible for the optical properties of different systems doped with RE and TM
ions. The synthesis of these materials has as its main goal the development of new lumines-
cent materials, specially those showing UC emission, and the study of the fundamental
physics, in particular the involved UC mechanisms, responsible for this luminescence.
First, luminescence properties of Y2 O3 and NaYF4 nanoparticles doped with Er3+ and
Yb3+ are investigated. The Er3+ -Yb3+ combination is frequently used in the development
of high efficiency UC materials [1], besides, NaYF4 is the most efficient host material up
to date for UC phosphors [2]. The underlying mechanisms in UC processes for Tb3+ -Yb3+
and Eu3+ -Yb3+ co-doped GGG and YAG nanocrystals are also analyzed. Then, green
UC Mn2+ luminescence in Mn2+ -Yb3+ co-doped LMA powders is demonstrated, as far as
we know, for the first time up to 650 K. In the last section, the temperature and pressure
dependence of the Cr3+ transitions in GGG nanoparticles is studied.
5.2 Rare-earth ions doped nanoparticles
UC materials, in which visible light is generated upon IR excitation, have attracted sig-
nificant attention for advanced applications such as solid state lasers [3], [4], materials
73
for biological applications like bio-labeling, drug delivery, or diagnostics [5], [6], UC phos-
phors, IR quantum counter detectors, or efficiency improvement of bifacial solar cells [7].
UC laser systems show several advantages over direct UV excitation; IR excitation re-
duces the photo-ionization induced degradation of hosts, it does not need high excitation
wavelength stability, and the output wavelength is not restricted to a given harmonic [3].
Recently, Prasad et al. used upconverting nanophosphors for in vitro and in vivo pho-
toluminescence bio-imaging, providing deeper light penetration into the biological tissues
which are transparent in the 750-1000 nm range. The advantage of NIR excitation is the
reduction of the background autofluorescence of the tissues [6].
The vast majority of UC studies investigated up to date involve insulating materials

doped with RE ions. The energy emission in semiconductors is known to be very sensitive
to the system dimensionality and to the particle size [8]. However, due to the localized
character of the f electrons, the emission energies associated to f−f transitions in RE ions
are mainly independent of the nanocrystals size. On the other hand, the emission intensity
and lifetime may change when particle size decreases due to the presence of organic
impurities on the nanoparticles surface. The emission intensity can also be modified
by the combination of RE ions and by changing the doping concentrations. It is well-
known that Yb3+ is an excellent UC sensitizer for Er3+ , Tm3+ , Pr3+ or Ho3+ ions [1], [9].
All these ions present intermediate states almost resonant with the 2 F7/2 → 2 F5/2 Yb3+
transition. The doubly-doped UC systems containing Er3+ and Yb3+ present the highest
UC efficiencies [2]. On the contrary, the situation is completely different for RE ions like
Tb3+ or Eu3+ which have no intermediate levels resonant with Yb3+ . Tb3+ and Eu3+ are
attractive as emitting ions because of their high quantum efficiency related to the large
energy gap between the emitting states and the low lying 7 FJ (J = 0, 1, . . . , 7) excited
states.
The ability of Yb3+ to induce UC is based on the high oscillator strength of the unique
f−f transition, 2 F7/2 → 2 F5/2 , which is located in the NIR, just in the range of cheap and
high power LD. Yb3+ has no higher excited states and it is transparent in the visible
region. The UC processes involving resonant ions, like in Er3+ -Yb3+ co-doped materials,
can be ascribed to GSA/ESA and/or GSA/ETU mechanisms. Since the Tb3+ -Yb3+ UC
system was introduced in 1969, many studies have been devoted to these two ions [10], [11],
[12], but the fact of getting UC emission in nanoparticles is not so common and always
noticeable. However, as far as we know, there are only few examples of UC luminescence
in Eu3+ -Yb3+ systems in the literature [13], [14]. The UC luminescence in Tb3+ -Yb3+
or Eu3+ -Yb3+ systems is assigned to a cooperative sensitization mechanism or GSA/ESA
processes in dimers.
In this work, nanocrystals of Y2 O3 : Er3+ , Yb3+ , NaYF4 : Er3+ , Yb3+ as well as
Tb3+ -Yb3+ and Eu3+ -Yb3+ co-doped GGG and YAG nanoparticles have been prepared
by using different synthesis methods. The potential applications of these nanocrystalline
materials go through the possibility of reaching the same properties observed in bulk but
in systems with dimensions in the range of nanometers. Silica is highly bio-compatible and
its surface chemistry is well documented for biological interactions; in this sense, coating
with SiO2 would be the first step in the functionalization of the obtained nanoparticles.
XRD, TEM, as well as RT luminescence, excitation and lifetime measurements have been
carried out in these samples. The most relevant structural and spectroscopic results in
order to characterize them and to identify the UC mechanisms in each case are shown in
this section.

Within inorganic materials, Y2 O3 is one of the most widely studied systems for accept-
ing trivalent impurities. It crystallizes in the cubic system, showing a bixbyte structure
((Fe,Mn)2 O3 ) with Ia3 space group (Fig. 5.1). The unit cell contains 16 molecules and
the bulk cell parameter is a0 =10.604 Å [15]. Y3+ ions are accommodated in 32 sites in the
unit cell; 24 sites with point group symmetry C2 and 8 sites with C3i symmetry. RE ions
substitute Y3+ ions and have been found to be randomly distributed in both sites [16].
The C3i site has associated a center of inversion, therefore, f-f electric dipole transitions
are much weaker than those related to the C2 site. The maximum phonon energy in bulk
yttria is 600 cm−1 [17].
In this work, Y2 O3 : Er3+ , Yb3+ nanoparticles prepared by two methods, ball milling
Figure 5.1: Y2 O3 cubic structure.
Figure 5.2: XRD patterns of Y2 O3 : 2%Er3+ , 1%Yb3+ nanoparticles prepared by ball milling for 2 and
180 hours (a), and calculated pattern for cubic Y2 O3 (a0 =10.604 Å) (b).
(top-down) and combustion (bottom-up) are studied. Both synthesis methods produce
nanoparticles of Y2 O3 cubic phase as it can be seen from XRD pattern. Figure 5.2 shows
the XRD patterns of the Y2 O3 : 2%Er3+ , 1%Yb3+ nanocrystalline samples prepared in the
planetary ball mill for 2 and 180 hours milling time according to the parameters defined in
Section 3.2. Considering the most intense peak in the XRD diagram for different grinding
times (Fig. 5.3), it can be observed that when grinding time increases, the peaks intensity
diminishes while their width increases. This broadening is due to both smaller particle
size and strains generated during the milling.
According to the Williamson-Hall equation (eq. 4.1), the average crystallite size can
Figure 5.3: Y2 O3 : 2%Er3+ , 1%Yb3+ XRD peak at 2θ=29.3◦ for different grinding times.
Figure 5.4: Average crystal size, L, for the Y2 O3 : 2%Er3+ , 1%Yb3+ ball milling sample as a function
of milling time.
be estimated from XRD diagrams. Considering the dependence of particle size on time
(Fig. 5.4), we realize that for times longer than 35 hours a steady-state particle size of
25 nm is reached. This limit value is a characteristic feature of each kind of material and
grinding process. We can conclude then, that the optimal milling time for doped Y2 O3
samples is around 50 hours.
Er3+ and Yb3+ co-doped Y2 O3 nanoparticles have also been prepared following the
glycine-nitrate solution combustion synthesis (reaction 3.4). Figure 5.5 shows the XRD
patterns obtained for a sample prepared by this method before and after it was treated at
500 ◦ C. It can be seen from the peaks broadening that in both cases the particle size is in
Figure 5.5: XRD patterns of Y2 O3 : 2%Er3+ , 1%Yb3+ nanocrystals prepared by combustion before
and after firing at 500 ◦ C (a), and calculated pattern for cubic Y2 O3 (Ia3 space group) (b).
Figure 5.6: Fitting of the experimental peak broadening to the Williamson-Hall equation for Y2 O3 :
2%Er3+ , 1%Yb3+ nanoparticles prepared by combustion and fired at 500 ◦ C for 1 hour.
the nanometer range. The fitting of the experimental peak broadening to the Williamson-
Hall equation for Y2 O3 : 2%Er3+ , 1%Yb3+ samples prepared by combustion and calcined
(Fig. 5.6) is given by Br cos θ = 0.01053 + 0.01407 sin θ. The average particle size of the
nanoparticles obtained by combustion is about 10 nm before calcination and it increases
to 15 nm after firing at 500 ◦ C for 1 hour.
The presence of CO2− −
3 and OH anionic groups on the nanoparticles surface is known
to be one of the main reasons for the diminishing of the emission intensity as well as for
the reduction of the emission lifetime. These ions contribute with high energy phonons
increasing the non-radiative relaxation processes via multiphonon relaxation according to
Figure 5.7: IR spectra of nanocrystalline Y2 O3 : 2%Er3+ , 1%Yb3+ prepared by combustion (up) and
ball milling (down). Sequential heat treatments have been carried out on combustion samples: 500 ◦ C
for 2 hours, and 1000 ◦ C for 5 h.
the gap’s law [17]. Moreover, this fact reveals surface effects in the luminescence properties
of nanocrystalline materials. The adsorption of these anionic groups is related to the
precursors used in the combustion synthesis but this cannot be the only reason, because
they also appear in the samples obtained by ball milling, where the starting materials are
free of these ions. In order to reduce surface contamination, different thermal treatments
have been performed on the nanocrystalline samples.
Figure 5.7 shows the medium infrared (MIR) spectra of Y2 O3 : 2%Er3+ , 1%Yb3+
prepared by both combustion and ball milling. All spectra show bands at approximately
1500 and 3400 cm−1 indicating the presence of CO2− −
3 and OH ions on the nanoparticles
surface. The intensity of these bands is reduced as calcination time and temperature
increase. The best nanoparticles were obtained after calcination at 1000 ◦ C for 5 hours.
Nevertheless, several hours after the heat treatment, the nanocrystalline samples were
able to adsorb again these anions.
Samples must be fired in order to reduce surface contamination but longer thermal
treatments at higher temperature sinterize the powders forming larger nanoparticles. Fig-
ure 5.8 presents the particle size dependence on calcination temperature, TC , and time,
tC , for Y2 O3 : 2%Er3+ , 1%Yb3+ combustion sample. It is observed that nanocrystals size
increases for longer firing times and higher temperatures.
Figure 5.8: Variation of Y2 O3 : 2%Er3+ , 1%Yb3+ nanoparticles size, L, as a function of the calcination
temperature (a), and time (b).
Figure 5.9: TEM image of Y2 O3 : 2%Er3+ , 1%Yb3+ prepared by combustion after SiO2 coating. In
the HRTEM images (b) and (c), the observed interplanar distances (0.33 nm) corresponds to the (222)
crystal plane.
Bearing in mind both effects, surface contamination and particle size, all Y2 O3 samples
studied in this work were fired at 900 ◦ C for an hour, and closed in shielded capillaries.
This covering was performed in order to avoid further contamination which produces a
deterioration of their optical properties. Some of the samples were also SiO2 coated. A
final average diameter of 50 nm has been estimated from the peak width of the XRD (see
Fig. 5.8). This value is in very good agreement with TEM results shown in figure 5.9.
TEM images of coated Y2 O3 nanoparticles show a SiO2 shell thickness of about 5-10 nm.
Optical properties
Optical properties of Y2 O3 : 2%Er3+ , 1%Yb3+ and Y2 O3 : 2%Er3+ , 20%Yb3+ nanoparticles

prepared by high energy ball milling and combustion synthesis are analyzed.
Figure 5.10 shows the RT diffuse reflectance spectrum of Y2 O3 : 2%Er3+ , 20%Yb3+

prepared by ball milling. The absorption peaks correspond to Er3+ transitions from
the ground state, 4 I15/2 , to different excited states, and to the 2 F7/2 → 2 F5/2 Yb3+
transition.
Figure 5.10: RT diffuse reflectance spectrum of Y2 O3 : 2%Er3+ , 20%Yb3+ prepared by ball milling.
Figures 5.11 and 5.12 compare the RT luminescence spectra of Y2 O3 doped with
different Er3+ and Yb3+ concentrations, and prepared by ball milling and combus-
tion, upon direct Er3+ excitation at 26596 cm−1 , and IR excitation at 10256 cm−1 .
Figure 5.11: RT emission of nanocrystalline Y2 O3 co-doped with different concentrations of Er3+

and Yb3+ upon excitation at 26596 cm−1 . Measurements were performed under the same experimental
conditions.
Figure 5.12: RT UC luminescence of Y2 O3 : Er3+ , Yb3+ nanoparticles with different RE concentrations

upon excitation at 10256 cm−1 . Measurements were performed under the same experimental conditions.
Figure 5.13: RT UC luminescence of Y2 O3 : 2%Er3+ , 20%Yb3+ nanoparticles prepared by combustion,

upon excitation at 10256 cm−1 , with and without SiO2 coating.
Emission peaks around 15000 and 18000 cm−1 are assigned, respectively, to the
transitions from the 4 F9/2 and the thermalized (2 H11/2 , 4 S3/2 ) excited states, to the
4
I15/2 ground state of Er3+ ions (Fig. 5.14). Figure 5.13 shows the UC emission
spectra of Y2 O3 : 2%Er3+ , 20%Yb3+ nanoparticles prepared by combustion with
and without SiO2 . The total intensity as well as the intensity ratio between red and
green emissions in the spectra of the nanoparticles with and without SiO2 coating
are exactly the same.
From the spectra in Fig. 5.11, we conclude that samples prepared by ball milling
have higher luminescence intensities than combustion samples. This is due to the
presence of CO2− −
3 and OH groups which is more important in combustion samples,
even after thermal treatment. A salient feature is that in the 2%Er3+ , 1%Yb3+
samples, the intensity of the green emission is higher than the red one. On the
contrary, those samples doped with 20%Yb3+ have a higher red to green intensity
ratio. This reduction in the green luminescence is related to the (2 H11/2 , 4 S3/2 ) +
2
F7/2 → 4 I11/2 + 2 F5/2 CR process between Er3+ and Yb3+ ions, whose probability
increases with increasing Yb3+ concentration, and is responsible for the additional
(2 H11/2 ,4 S3/2 ) depopulation channel (Fig. 5.14).
Figure 5.12 shows the UC luminescence spectra obtained at RT for Y2 O3 : 2%Er3+ ,

Figure 5.14: Energy level scheme of Er3+ and Yb3+ ions and the proposed UC mechanisms. Green
and red luminescence in Y2 O3 : Er3+ , Yb3+ samples are depicted. Cross relaxation is responsible for the
reduction of green emission. GSA/ESA or GSA/ETU from 4 I13/2 are responsible for the enhanced red
emission upon IR excitation.
1%Yb3+ and Y2 O3 : 2%Er3+ , 20%Yb3+ prepared by mechano-chemical and combus-

tion synthesis upon IR LD excitation centered at 10256 cm−1 . The broad excitation
(∆λ ∼ 10 nm) is able to excite both 2 F7/2 → 2 F5/2 Yb3+ and 4 I15/2 → 4 I11/2 Er3+
transitions. The same bands around 15000 and 18000 cm−1 are observed, and ball
milling nanocrystals are once again more efficient than combustion ones. The red UC
emission, 4 F9/2 → 4 I15/2 , is more intense than the green UC luminescence, (2 H11/2 ,
4
S3/2 ) → 4 I15/2 in all cases. For 20%Yb3+ -doped nanoparticles the green emission
is not detected by the naked eye. The enhancement of red to green emission upon
IR excitation can be explained with the following process: after IR excitation, both
2
F5/2 Yb3+ and 4 I11/2 Er3+ multiplets are populated. Yb3+ excitation is transferred
to the 4 I11/2 Er3+ level by energy transfer (GSA/ETU). Then, most of the ions decay
non-radiatively to the 4 I13/2 level, and Er3+ ions can reach the 4 F9/2 state after the
absorption of a second IR photon, ESA, or after another ETU process (Fig. 5.14).
In the absence of the 4 I11/2 → 4 I13/2 non-radiative process, a second photon could
reach the (2 H11/2 , 4 S3/2 ) green emitting states. This additional mechanism which
seems to be more efficient than the (2 H11/2 , 4 S3/2 ) → 4 F9/2 decay, does not exist
upon visible excitation.
Temporal evolution of the red, 4 F9/2 → 4 I15/2 , and green, (2 H11/2 , 4 S3/2 ) → 4 I15/2 ,
Er3+ luminescence after pulsed excitation with the OPO at 20492 cm−1 and 10256
cm−1 has been recorded.
Figure 5.15 shows the time dependence of the red and green Er3+ luminescence of
Y2 O3 : 2%Er3+ , 20%Yb3+ prepared in the planetary ball mill upon excitation at
20492 cm−1 . Experimental data obtained after visible excitation have been fitted
to a single exponential, I(t) = I0 exp( −t
τ
) for all samples. The best fit parameters
are shown in Table 5.1. Both emitting states, 4 S3/2 and 4 F9/2 , show longer lifetimes
in ball milling samples in comparison with combustion ones. This fact reveals that
the intensity diminishing observed for samples prepared by combustion is related
to non-radiative processes, which are more important in samples prepared following
the combustion reaction.
Figure 5.15: Temporal evolution of the RT red (a) and green (b) Er3+ luminescence of Y2 O3 : 2%Er3+ ,
20%Yb3+ prepared by ball milling upon excitation at 20492 cm−1 .
Figure 5.16 compares the temporal behavior of the red Er3+ UC luminescence after
IR excitation at 10256 cm−1 for Y2 O3 : 2%Er3+ , 20%Yb3+ prepared by ball milling
and combustion synthesis. A rise of the emission intensity followed by a decay is
detected in both cases. In this case, the temporal evolution of the Er3+ UC emission
has been fitted to a Vial’s type equation, I(t) = A exp( −t
B
)−C exp( −t
D
) [18]. B and D
represent the decay and rise of the transient, respectively; hence, B is essentially the
inverse of 4 S3/2 or 4 F9/2 states lifetime, and D is related to the ETU rate (WETU ) and
Table 5.1: Lifetimes of the 4 F9/2 and 4 S3/2 Er3+ levels in different Y2 O3 : Er3+ , Yb3+
samples detecting at 15129 cm−1 and 17730 cm−1 , respectively, after direct excitation in
the 4 F7/2 level at 20492 cm−1 .
Y2 O3 : Er3+ , Yb3+ Sample τ (4 F9/2 )/µs τ (4 S3/2 )/µs

2%Er3+ , 20%Yb3+ Combustion 2.2 ± 0.1 1.1 ± 0.1
3+ 3+
2%Er , 1%Yb Ball milling 17.6 ± 0.4 4.1 ± 0.2
3+ 3+
2%Er , 1%Yb Combustion 15.3 ± 0.6 2.8 ± 0.2
Figure 5.16: Temporal evolution of the RT 4 F9/2 Er3+ UC luminescence of Y2 O3 : 2%Er3+ , 20%Yb3+
prepared by ball milling (a) and combustion (b) upon excitation at 10256 cm−1 .
1
2
F5/2 , 4 I11/2 lifetimes, according to D = τ (2 F5/2 )
+ τ (4 I111/2 ) + WETU . The best fits for
red and green Er3+ UC luminescence are shown on Tables 5.2 and 5.3, respectively.
The rise of the luminescence intensity after IR excitation is observed in both red
and green UC emissions in all samples. This is a clear proof of an energy transfer
process and evidences the contribution of GSA/ETU mechanism to the total UC
luminescence. In some cases the rise is not starting from zero, this implies that there
is also a contribution of the Er3+ single ion GSA/ESA mechanism; that is, after IR
excitation, an excited Er3+ in the 4 I11/2 or 4 I13/2 level can absorb a second photon
and be promoted to the 4 F9/2 or (2 H11/2 , 4 S3/2 ) levels, respectively. Figure 5.17
shows the temporal evolution of the green Er3+ UC luminescence of Y2 O3 : 2%Er3+ ,
1%Yb3+ prepared by ball milling upon excitation at 10256 cm−1 . Taking into ac-
count that the lifetime of the 4 S3/2 state in the GSA/ESA process should be exactly
the same that after direct Er3+ excitation at 20492 cm−1 , it is possible to estimate
Table 5.2: Lifetimes of the 4 F9/2 Er3+ level in different Y2 O3 : Er3+ , Yb3+ samples mea-
(see text).
Y2 O3 : Er3+ , Yb3+ Sample B(4 F9/2 )/µs D(4 F9/2 )/µs

2%Er3+ , 20%Yb3+ Combustion 7.5 ± 0.4 1.7 ± 0.2
3+ 3+
2%Er , 1%Yb Ball milling 163 ± 5 1.7 ± 0.1
2%Er3+ , 1%Yb3+ Combustion 101 ± 7 1.7 ± 0.2
Table 5.3: Lifetimes of the 4 S3/2 Er3+ level in different Y2 O3 : Er3+ , Yb3+ samples mea-
(see text).
Y2 O3 : Er3+ , Yb3+ Sample B(4 S3/2 )/µs D(4 S3/2 )/µs

2%Er3+ , 20%Yb3+ Combustion 1.7 ± 0.2 0.50 ± 0.04
2%Er3+ , 1%Yb3+ Ball milling 122 ± 5 5.0 ± 0.2
2%Er3+ , 1%Yb3+ Combustion 80 ± 10 6.0 ± 0.5
Figure 5.17: Temporal evolution of Er3+ 4 S3/2 UC luminescence intensity in Y2 O3 : 2%Er3+ , 1%Yb3+
obtained by ball milling detecting at 17730 cm−1 upon excitation at 10256 cm−1 . The inset shows a
detail of the intensity rise. The shadow region corresponds to the GSA/ESA contribution taking the
intrinsic lifetime of the 4 S3/2 state after direct excitation.
the GSA/ESA and GSA/ETU contribution to the total intensity. Contributions

of 99% and 1% have been estimated for GSA/ETU and GSA/ESA mechanisms,
respectively [19].
The green (2 H11/2 , 4 S3/2 ) emission lifetimes for Y2 O3 : 2%Er3+ , 1%Yb3+ samples,
observed upon excitation at 10256 cm−1 , are much longer than those obtained for
samples doped with 20%Yb3+ . Even more, lifetimes after IR excitation are longer
than lifetimes after direct Er3+ excitation at 20492 cm−1 . This highlights the energy
transfer in which Yb3+ ions, with a longer lifetime of about 1000 µs, can feed the
4
I11/2 Er3+ multiplet after IR excitation. The same trend is observed for the red
emission decay. The shorter lifetimes detected for samples doped with 20%Yb3+ are
related to a back-transfer from Er3+ to Yb3+ ions whose probability increases with
Yb3+ concentration (see Fig. 5.14).
Conclusions
Y2 O3 : Er3+ , Yb3+ nanocrystalline samples have been synthesized by two different meth-
ods, mechano-chemical synthesis and combustion reaction. A steady-state grain size of 25
nm is reached for the milled Y2 O3 samples while average sizes of 10-15 nm are estimated
for nanoparticles prepared by combustion. However, thermal treatments are required in
order to reduce surface contamination responsible for the luminescence quenching and fi-
nal sizes of ca. 50 nm are obtained for all samples. The nanocrystals have been protected
in capillaries for optical studies and in some cases SiO2 -coated (10 nm thickness). In both
cases, the optical properties are similar indicating the preservation of the nanocrystals
for further surface contamination. The presence of a SiO2 coating surface would allow
to functionalize the nanoparticles for different applications. The dependence of optical
properties on synthesis method or dopant concentration has been studied. The red to
green intensity ratio can be tuned by changing Yb3+ concentration. More intense UC
emission and longer lifetimes have been detected for ball milling samples making them
more suitable for applications. GSA/ETU has been established to be the main mechanism
responsible for the UC luminescence. An enhancement of red emission after IR excitation
has been identified.

Sodium yttrium fluoride, NaYF4 , can crystallize with two different phases, the hexagonal
β-phase (space group P − 6) and the cubic α-phase (space group F m3m). As it will be
discussed later, obtaining NaYF4 : Er3+ , Yb3+ nanoparticles in the pure β phase would
be an important step to get a promising efficient UC material.
Figure 5.18 shows the XRD pattern of NaYF4 nanoparticles co-doped with nominal
concentrations of 2%Er3+ , 20%Yb3+ , synthesized in a planetary ball mill according to the
reaction and the optimal parameters described in Section 3.2, and using different milling
times. It must be noted that when milling time increases, both intensity and width of the
peaks corresponding to the hexagonal phase increase and β-NaYF4 is obtained for times
longer than four hours. On the contrary, XRD peaks matching with cubic phase decrease
when milling time increases. Peaks which do not coincide with any of these phases are
related to the initial YF3 and disappear for times longer than four hours.
Figure 5.18: XRD patterns of NaYF4 : 2%Er3+ , 20%Yb3+ for different milling times (0.5, 2.5, 4.5 and
13 hours). The calculated patterns for hexagonal β-NaYF4 and cubic α-NaYF4 are also shown. Asterisks
indicate YF3 peaks.
The average nanoparticles size is deduced from XRD peaks broadening according to
the Williamson-Hall equation (eq. 4.1). The particle size estimated for NaYF4 : 2%Er3+ ,
20%Yb3+ is about 33 nm after 4 hours milling (Fig. 5.19(a)), and 36 nm after 13 hours
Figure 5.19: Particle size estimation for NaYF4 : 2%Er3+ , 20%Yb3+ nanoparticles prepared in the ball
mill after 4.5 hours (a) and 13 hours (b) milling time. The fitting Williamson-Hall equations are given by
Br cos θ = 0.00467 + 0.00727 sin θ (a) and Br cos θ = 0.00432 + 0.00599 sin θ (b) and the estimated sizes
are L=33 nm (a) and L=36 nm (b), respectively.
Figure 5.20: TEM image of NaYF4 : 2%Er3+ , 20%Yb3+ nanocrystals prepared by ball milling.
(Fig. 5.19(b)). Therefore, it can be seen that from 4 hours, further milling time does
not imply significant changes in particle size. Considering not only the hexagonal phase
formation with the lower cubic contribution, but also the resultant nanocrystals size, it
can be concluded that the optimum milling time for β-NaYF4 : 2%Er3+ , 20%Yb3+ is ca.
10 hours.
A TEM image of NaYF4 : 2%Er3+ , 20%Yb3+ nanocrystals prepared by ball milling

after 13 hours grinding is shown in Fig. 5.20. An average size of ca. 50 nm is estimated,
this value is slightly larger than the one obtained from XRD patterns broadening (36 nm).
Particles are often aggregated forming bigger clusters.
In order to analyze the efficiency of these nanoparticles prepared by ball milling as UC

phosphor material, their optical properties have been compared with the experimental
Figure 5.21: XRD patterns of NaYF4 : 2%Er3+ , 20%Yb3+ bulk and nanoparticles obtained for different
grinding times (1, 2 and 4 hours). The calculated patterns for hexagonal β-NaYF4 and cubic α-NaYF4
are also shown.
results obtained for the bulk β-NaYF4 : 2%Er3+ , 20%Yb3+ synthesized by Krämer [2].
Moreover, NaYF4 : 2%Er3+ , 20%Yb3+ nanoparticles have also been obtained by milling
this bulk β-NaYF4 crystals of 2-5 µm in size in the planetary ball mill. This process
has been carried out using ZrO2 vials and balls, a ball to powder mass ratio of 20:1, an
angular velocity of 300 rpm and stopping 5 minutes every 30 minutes of milling.
Figure 5.21 shows the XRD patterns of the as prepared NaYF4 : 2%Er3+ , 20%Yb3+
nanoparticles for 1, 2 and 4 hour milling and the bulk material. It is observed that the
bulk crystallizes only in the hexagonal system, however, when grinding time increases,
the hexagonal peaks intensity decreases while peaks corresponding to the cubic phase
increase. From hexagonal peaks broadening an average size of 30 nm is obtained for both
1 and 4 hours milling time.
Optical properties
Hexagonal β-NaYF4 doped with Er3+ -Yb3+ and Tm3+ -Yb3+ are up to now, the most
efficient bulk materials for green and blue UC luminescence, respectively [2]. It is an
order of magnitude more efficient than the cubic phase. This can be ascribed to the
low vibrational energies in the hexagonal lattice, phonon cutoff around 350 cm−1 , which
reduces non-radiative multiphonon relaxation processes [20]. Besides, the high emission
efficiency has also been attributed to the structural disorder of the lattice and the presence
of multi-sites for Na+ and RE3+ ions [21], [22]. Both the doping ratio and the phase purity
determine the actual UC efficiency. NaYF4 : 2%Er3+ , 20%Yb3+ colloidal nanocrystals
showed a reduction in the UC efficiency by a factor of 102 -103 [1]. Half of this diminution
was believed to be due to the fact that the nanoparticles crystallize in the α phase, and
the other half to the presence of OH− impurities on the particles surface.
• Synthesis method dependence.
In this work we compare optical properties and UC luminescence efficiency of

NaYF4 : 2%Er3+ , 20%Yb3+ β-bulk and nanoparticles prepared by milling the bulk
hexagonal micropowders, and by the mechano-chemical reaction of the fluorides
described in Section 3.2.
Figure 5.22: RT luminescence spectra of NaYF4 : 2%Er3+ , 20%Yb3+ bulk and nanoparticles (30
nm) upon Er3+ excitation at 26596 cm−1 . Measurements were performed under the same experimental
conditions.
Figures 5.22 and 5.23 show the RT luminescence spectra of bulk β-NaYF4 : 2%Er3+ ,
20%Yb3+ and nanoparticles obtained by milling the bulk powders for 1, 2 and 4
hours (average size of 30 nm) upon direct Er3+ excitation at 26596 cm−1 and af-
ter IR excitation at 10256 cm−1 . Emission bands around 15000 and 18000 cm−1
Figure 5.23: RT UC emission spectra of NaYF4 : 2%Er3+ , 20%Yb3+ bulk and nanoparticles (30
nm) upon IR excitation at 10256 cm−1 . Measurements were carried out under the same experimental
conditions.
detected in both cases are assigned to the red 4 F9/2 → 4 I15/2 and green (2 H11/2 ,
4
S3/2 ) → 4 I15/2 Er3+ transitions, respectively. It is worth mentioning that after vis-
ible excitation into the 4 G11/2 Er3+ level, red emission is more intense than green
luminescence and when milling time increases from 1 to 4 hours, the red to green
intensity ratio decreases by a factor of two. On the other hand, an enhancement of
green emission upon IR excitation is detected for all samples. An evident reduction
in the overall luminescence intensity is observed upon both visible and IR excitation
for longer milling times. The intensity decrease is more noticeable in the UC lumi-
nescence and a reduction in the UC efficiency by a factor of 50 and 103 is estimated
for NaYF4 : 2%Er3+ , 20%Yb3+ nanocrystals obtained after 1 and 4 hours milling,
respectively. The low efficiency observed in nanoparticles is related to two different
factors regarding the milling process. First, the cubic phase emergence which is less
efficient than the hexagonal phase [2]. Second, the particles surface contamination
with the vials and balls which is known to be one of the main problems in mechanical
milling synthesis [23].
Figure 5.24 compares the RT UC luminescence spectra of NaYF4 : 2%Er3+ , 20%Yb3+

obtained using the two described procedures; by milling the hexagonal bulk doped
Figure 5.24: RT UC luminescence spectra of NaYF4 : 2%Er3+ , 20%Yb3+ prepared by milling the
hexagonal bulk (red) and by the mechano-chemical reaction of the fluorides (black) upon IR excitation
at 10256 −1 . Notice the scale factor.
NaYF4 for 1 hour, and by the mechano-chemical reaction of the fluorides (NaF, YF3 ,
YbF3 , ErF3 ) according to equation 3.2 for ten hours. In both cases, nanoparticles
size is around 30-35 nm. It must be pointed out that the sample prepared from the
fluorides reaction is much less efficient (notice the scale factor). The fact that in
this sample the red emission is five times more intense than the green luminescence
is also noteworthy. Since the impurity distribution during the mechano-chemical
reaction is difficult to control, it might be very heterogenous. Therefore, there
is a higher probability of finding nanoparticles with impurity concentration larger
than nominal. This situation makes non-radiative process, responsible for the lu-
minescence quenching, become more important in nanocrystals prepared by the
mechano-chemical reaction.
The RT green (2 H11/2 , 4 S3/2 ) → 4 I15/2 and red 4 F9/2 → 4 I15/2 Er3+ UC luminescence
intensity versus the excitation power density at 10250 cm−1 for NaYF4 : 2%Er3+ ,
20%Yb3+ nanoparticles obtained after 1 hour milling is plotted on a double logarith-
mic scale in Fig. 5.25. A quadratic power dependence is observed below 1 W·cm−2
Figure 5.25: Excitation power dependence of the Er3+ green (a) and red (b) UC emission upon exci-
tation at 10250 cm−1 for NaYF4 : 2%Er3+ , 20%Yb3+ milled for 1 hour.
for both emissions, which is the typical behavior of a two-photon excitation process
in the low-power regime. Hence, this behavior represents the experimental confir-
mation of a two-photon UC process. On the contrary, the slope decreases to 1.0 and
1.2 for green and red luminescence, respectively, at higher excitation power densities
as explained by Pollnau et al. [24] and Suyver et al. [25].
RT temporal evolution of the red and green Er3+ UC luminescence in NaYF4 :

2%Er3+ , 20%Yb3+ bulk and nanoparticles prepared by milling the bulk powders
and by the mechano-chemical reaction of the fluorides have been measured upon
visible and IR excitation (data not shown). The decays measured upon visible exci-
tation at 20492 cm−1 have been fitted to a single exponential. In the experimental
data obtained upon IR excitation at 10246 cm−1 a rise followed by a decay is de-
tected, similarly to the case of Y2 O3 : Er3+ , Yb3+ nanoparticles described previously
in this section. The UC emission decay have been fitted to a Vial’s type equation
[18]. The best fitting values are enclosed on tables 5.4 and 5.5, B and D being the
decay and rise, respectively.
Comparing the temporal evolution of the luminescence intensity for NaYF4 : 2%Er3+ ,
20%Yb3+ bulk and nanoparticles, it can be seen that lifetimes obtained for bulk are
systematically longer than those obtained for nanoparticles, for both red and green
luminescence upon visible and IR excitation. This lifetime decrease associated to
the size diminution is partially related to non-radiative process since, as we have
Table 5.4: Lifetimes of the 4 F9/2 Er3+ level in different NaYF4 : 2%Er3+ , 20%Yb3+ samples

(Eexc =20492 cm−1 ) (Eexc =10246 cm−1 ) (Eexc =10246 cm−1 )
Bulk 385 ± 5 655 ± 5 60 ± 2
1 hour milled 145 ± 5 580 ± 10 230 ± 5
9.5 hours milled 54 ± 4 78 ± 5 5.8 ± 0.5
Table 5.5: Lifetimes of the 4 S3/2 Er3+ level in different NaYF4 : 2%Er3+ , 20%Yb3+ samples

(Eexc =20492 cm ) (Eexc =10246 cm ) (Eexc =10246 cm−1 )
−1 −1
Bulk 116 ± 2 316 ± 5 42 ± 2

1 hour milled 33 ± 1 177 ± 6 60 ± 3
9.5 hours milled 11.5 ± 0.5 23.5 ± 0.8 2.3 ± 0.2
seen, it is accompanied by an intensity decrease. However, the intensity decrease

is notably more important than the lifetime reduction. As an example, in NaYF4 :
2%Er3+ , 20%Yb3+ milled for 1 hour the green UC luminescence intensity diminishes
by a factor of 50 while the lifetime is only reduced to half its value. This additional
intensity reduction should be related to the cubic phase appearance.
Conclusions
NaYF4 : 2%Er3+ , 20%Yb3+ nanocrystals have been prepared by two different methods,
and particles of around 30 nm in size are obtained in all cases. The sample obtained
after milling the bulk for 1 hour shows the best UC luminescence efficiency but it is still
much less efficient than bulk. This is related to the emergence of cubic α-NaYF4 , and to
surface contamination of the nanocrystals during the milling process, associated to the
larger surface-to-volume ratio in nanoparticles.
5.2.3 Tb3+ or Eu3+ and Yb3+ co-doped Gd3 Ga5 O12 and Y3 Al5 O12
Gadolinium gallium garnet, GGG, and yttrium aluminum garnet, YAG, are suitable mate-
rials as host for luminescent trivalent lanthanide ions. They belong to the cubic crystalline
system, having a garnet structure, with Ia3d space group (Fig. 5.26). The bulk cell pa-
rameters are a0 =12.376 Å and a0 =12.010 Å for GGG and YAG, respectively. For both
systems, the garnet structure is composed of a 24(c) dodecahedral site (D2 point symme-
try) for Gd3+ or Y3+ depending on the case with a coordination number 8, and two sites
for Ga3+ and Al3+ , respectively, a 16(a) octahedral site with a coordination number 6,
and a 24(d) tetrahedral site of coordination number 4. O2− ions occupy the 96(h) sites
with each one being a member of two dodecahedra, one octahedron and one tetrahedron.
The garnet structure can be viewed as interconnected dodecahedra, octahedra and tetra-
hedra with shared O atoms at the corners of the polyhedra. Taking into account the ionic
radii, the doping RE3+ impurities are expected to enter into the Gd3+ or Y3+ sites which
possess D2 symmetry [26].
Figure 5.26: GGG (left) and YAG (right) cubic structures. In GGG structure the two Ga3+ sites
with octahedral and tetrahedral coordination is shown, while the Y3+ site with coordination number 8
is depicted in YAG structure.
GGG and YAG nanocrystalline powders co-doped with nominal concentrations of

2%Tb3+ -5%Yb3+ , 2%Eu3+ -5%Yb3+ , and 2%Eu3+ -1%Er3+ have been prepared by the sol-
gel Pechini’s method described in Section 3.4. It is worth mentioning that lanthanides
replace Gd3+ or Y3+ ions, hence, the amount of RE3+ doping ions (1%, 2% or 5% mol)
is with respect to Gd3+ or Y3+ . All the prepared samples have been analyzed using the
XRD technique.
Figure 5.27 shows the XRD pattern of GGG: 2%Eu3+ , 5%Yb3+ . The Rietveld re-
finement of the GGG nanopowders XRD pattern is consistent with a cubic garnet single
phase (space group Ia3d, a=12.391 Å) but contains small traces (up to 5%) of a second
phase, Gd3 GaO6 . The obtained lattice constant is similar to that found for an Eu3+
doped GGG nanocrystalline sample prepared by the same Pechini procedure [27]. From
Rietveld refinement and considering XRD peaks broadening, the average size of the crys-
tallite grains has been determined to be ca. 30 nm. The Eu3+ -Er3+ doped GGG samples
are also cubic garnet single phase apart from a small trace contamination of Gd3 GaO6
phase, as found for the Eu3+ -Yb3+ doped samples.
Figure 5.27: XRD pattern of GGG: 2%Eu3+ , 5%Yb3+ prepared following Pechini’s method (a) and
Rietveld fit (b). The asterisk indicates the main peak of the Gd3 GaO6 impurity phase.
Figure 5.28 presents the XRD pattern of YAG: 2%Eu3+ , 5%Yb3+ together with the
Rietveld fit. Both the Eu3+ -Yb3+ and Eu3+ -Er3+ doped nanocrystalline YAG samples
crystallize only in the cubic garnet phase (space group Ia3d, a=12.031 Å) without evi-
dence of contamination from other phases. The corresponding size of the nanoparticles
obtained from Rietveld fitting is around 40 nm. There is no evidence of extra peaks due
to phase segregation of the doping components.
TEM images of GGG and YAG nanocrystalline samples (see Fig. 5.29) show that the
samples are made of particles of different shapes and sizes. Most of them are spherical
with sizes in the 30-50 nm range but in some cases particles are aggregated forming larger
clusters.
Figure 5.28: XRD pattern of YAG: 2%Eu3+ , 5%Yb3+ prepared by the sol-gel Pechini’s method (a)
and Rietveld fit (b).
Figure 5.29: TEM images of GGG: 2%Tb3+ , 5%Yb3+ prepared by the sol-gel Pechini’s method. In
the HRTEM image, the observed interplanar distances (0.30 nm) corresponds to the (012) crystal plane.
Figure 5.30 shows IR absorption spectra of both host materials, GGG and YAG, be-
tween 300 and 1300 cm−1 . Analyzing this frequency range, information about vibrational
energy of each matrix can be obtained. The higher energy absorption band is observed
at around 680 cm−1 for the GGG, while it is shifted to higher energies and appears at
around 800 cm−1 in YAG. The highest energy phonons belong to the Ga-O and Gd-O
Figure 5.30: IR absorption spectra of GGG: 2%Eu3+ , 5%Yb3+ and YAG: 2%Eu3+ , 5%Yb3+ synthesized
by Pechini’s method. Nanopowders were diluted in KBr to perform these measurements.
or Y-O and Al-O stretching vibration modes for GGG and YAG, respectively [28]. Ac-
cording to the gap’s law, multiphonon relaxation becomes much more probable for higher
phonon energies. Therefore, non-radiative processes, responsible for the diminishing of
luminescence, are more likely to happen in YAG than in GGG.
Optical properties
GGG and YAG are host materials with interesting optical properties [29]. Average phonon
energy in sesquioxides is quite low for oxide materials, which is especially beneficial for
optical applications. YAG: Nd3+ is one of the most important materials used for solid-
state lasers. GGG has also shown to have several advantages over other laser materials
[30]. A detailed investigation on the spectroscopy and excited state dynamics at RT has
been carried out on Tb3+ -Yb3+ and Eu3+ -Yb3+ co-doped GGG and YAG.
• Tb3+ -Yb3+ System.
Figure 5.31 shows Tb3+ excitation spectra in both GGG and YAG co-doped with
2%Tb3+ , 5%Yb3+ nanoparticles. In addition to the f−f Tb3+ transitions from the
ground state 7 F6 to different excited states plotted in the Dieke diagram, the excita-
tion spectrum of YAG: 2%Tb3+ , 5%Yb3+ shows two intense bands at around 31060
and 36100 cm−1 . We assign these transitions to the spin-allowed and spin-forbidden
4f−5dt transitions of Tb3+ , in agreement with the data reported by Dorenbos for
Figure 5.31: RT excitation spectra of GGG: 2%Tb3+ , 5%Yb3+ and YAG: 2%Tb3+ , 5%Yb3+ recording
Tb3+ luminescence at 18420 cm−1 .
the same compound (36500 and 30860 cm−1 ) [31], [32]. The high energy band is
more than one order of magnitude stronger than the low energy band. Surprisingly,
these bands are not detected in the case of GGG: 2%Tb3+ , 5%Yb3+ pointing out a
different crystal-field strength on the Tb3+ site.
Figure 5.32 presents Tb3+ 5 D4 → 7 FJ emission upon UV excitation in the 5dt state
at 37040 cm−1 for GGG: 2%Tb3+ , 5%Yb3+ and YAG: 2%Tb3+ , 5%Yb3+ nanocrys-
talline samples. As expected from excitation spectra, Tb3+ luminescence is one
order of magnitude more intense in YAG than in GGG.
Figure 5.33 compares the RT luminescence spectra of nanocrystalline GGG: 2%Tb3+ ,

5%Yb3+ and YAG: 2%Tb3+ , 5%Yb3+ upon visible and IR excitation. The first is
obtained exciting directly the Tb3+ ions at 37040 cm−1 while the UC luminescence
is observed after the excitation of Yb3+ ions at 10250 cm−1 . In all cases, the emis-
sion bands observed in the red-blue region (15000-21000 cm−1 ) are assigned to the
transitions from the 5 D4 multiplet to lower energy 7 FJ states of Tb3+ ions (see Fig.
5.34). Yb3+ pairs luminescence is not detected in our samples.
The presence of the parity allowed 4f−5dt transitions in the near-UV region in the
case of YAG: 2%Tb3+ , 5%Yb3+ , and additionally, the demonstration of Tb3+ -Yb3+
UC luminescence, opens the possibility to use this system also as a downconversion
Figure 5.32: RT luminescence spectra of GGG: 2%Tb3+ , 5%Yb3+ and YAG: 2%Tb3+ , 5%Yb3+ upon
Figure 5.33: RT luminescence spectra of GGG: 2%Tb3+ , 5%Yb3+ exciting at 37040 cm−1 (a) and
10250 cm−1 (b). RT emission spectra of YAG: 2%Tb3+ , 5%Yb3+ exciting at 37040 cm−1 (c) and 10250
cm−1 (d). * is an artifact and represents the laser at twice the excitation frequency.
Figure 5.34: Energy level scheme of Tb3+ and Yb3+ ions with the 5 D4 →7 FJ and the 5 D3 →7 FJ
Tb3+ luminescent transitions and the cooperative sensitization mechanism.
phosphor. In this process, one Tb3+ ion would generate two Yb3+ ions in the excited
state [33]. This could be applied for the improvement of solar cells efficiency. This is
particularly important taking into account that the 4f−5dt bands in YAG: 2%Tb3+ ,
5%Yb3+ are located close to band gap of clear glasses usually integrated in front of
the solar cell panels as protectors.
As we have seen, the excitation spectrum of Tb3+ in YAG (Fig. 5.35(a)) consists
of two strong bands, at 36100 cm−1 and 31060 cm−1 , compared to the intensity
of the f−f transitions. Excitation on the 4f−5dt bands is followed by a fast non-
radiative relaxation to the 5 D3 and 5 D4 excited states, from where emission takes
place (Fig. 5.35(b)). Although green 5 D4 → 7 FJ Tb3+ UC luminescence has been
widely investigated, there are only few examples of blue Tb3+ UC emission from
the 5 D3 multiplet in the literature [34], [35]. Luminescence from this state between
22000 and 27000 cm−1 has been observed after IR excitation of YAG: 2%Tb3+ ,
5%Yb3+ (Fig. 5.35(c)).
The only way to access the 5 D3 excited state is via a three photon process (see
Figure 5.35: RT excitation spectrum of YAG: 2%Tb3+ , 5%Yb3+ (a), RT luminescence spectrum after
excitation at 37037 cm−1 (b), and RT emission spectrum after excitation at 10250 cm−1 (c). Note the
scale factors.
Fig. 5.34). In order to prove this assumption, the power dependence of the emitted
photons of the 5 D4 → 7 FJ and 5 D3 → 7 FJ Tb3+ transitions upon IR excitation has
been measured (Fig. 5.36). Slopes of 2.6 and 3.3 are obtained for 5 D4 → 7 FJ and 5 D3
→ 7 FJ Tb3+ emissions, respectively. With this result, we can propose that a three
photon process is also involved in the green 5 D4 → 7 FJ luminescence, indicating
the relevance of a non-radiative relaxation from the 5 D3 to 5 D4 states. Such a three
photon process for the green UC Tb3+ emission was previously observed [36]. This
is in agreement with a possible CR process populating the 5 D4 from the 5 D3 state
[37].
• Eu3+ -Yb3+ System.
Figure 5.37 presents Eu3+ luminescence upon excitation at 25450 cm−1 for GGG:
2%Eu3+ , 5%Yb3+ and YAG: 2%Eu3+ , 5%Yb3+ nanoparticles under the same exper-
imental conditions. It can be observed that emission intensity is of the same order of
magnitude for both cases. Emission features are different for Eu3+ -Yb3+ co-doped
5
GGG and YAG nanoparticles. D0 → 7 F6,5 Eu3+ emission peaks are shifted to
Figure 5.36: Excitation power dependence of the Tb3+ 5 D4 → 7 FJ luminescence (a) and Tb3+ 5 D3 →
7
FJ emission (b) in YAG: 2%Tb3+ , 5%Yb3+ upon 10250 cm−1 excitation.
higher energies for GGG, 5 D0 → 7 F3,2 transitions have very diverse shape in GGG
and YAG, and 5 D0 → 7 F1 transitions show different energy separation between lines.
Figure 5.37: RT emission spectra of GGG: 2%Eu3+ , 5%Yb3+ and YAG: 2%Eu3+ , 5%Yb3+ upon
Figure 5.38 shows the Stokes and anti-Stokes RT luminescence spectra of nanocrys-
talline GGG: 2%Eu3+ , 5%Yb3+ and YAG: 2%Eu3+ , 5%Yb3+ . Eu3+ emission is
obtained upon direct excitation at 25450 cm−1 , whereas Eu3+ UC luminescence
is detected upon Yb3+ ions excitation at 10250 cm−1 . Unlike Yb3+ to Tb3+ UC
luminescence, there are only few examples of Yb3+ to Eu3+ UC emission in the
literature [13], [14], [38]. The bands observed in the spectra are assigned to 5 D0 →
7
FJ Eu3+ transitions (see Fig. 5.39). In the UC luminescence spectra (Fig. 5.38(b)
and 5.38(d)), emission bands in the 14500-15500 cm−1 range due to Er3+ impurities
are detected in addition to the Eu3+ f−f transitions. This fact was also observed by
Tanner [38] even with starting materials with purity of 99.999%. In order to under-
stand if Er3+ impurities play any role in the Yb3+ to Eu3+ UC luminescence process,
GGG: 2%Eu3+ , 1%Er3+ and YAG: 2%Eu3+ , 1%Er3+ nanocrystalline samples have
also been investigated.
Figure 5.38: RT luminescence spectra of GGG: 2%Eu3+ , 5%Yb3+ exciting at 25450 cm−1 (a) and
10250 cm−1 (b). RT emission spectra of YAG: 2%Eu3+ , 5%Yb3+ exciting at 25450 cm−1 (c) and 10250
cm−1 (d). The emission corresponding to Er3+ transitions are shown between arrows.
Figure 5.40 compares the UC luminescence of GGG: 2%Eu3+ , 1%Er3+ upon excita-
tion of Er3+ ions with a LD at 10250 cm−1 with that of GGG: 2%Eu3+ , 5%Yb3+ .
Er3+ excitation at 10250 cm−1 (Fig. 5.40 (a)) is able to induce UC luminescence
bands observed at around 15000 and 18000 cm−1 , assigned to the 4 F9/2 → 4 I15/2 ,
and (2 H11/2 , 4 S3/2 ) → 4 I15/2 Er3+ transitions, respectively, with the absence of Eu3+
emission. It is concluded that excitation at 10250 cm−1 induces Er3+ luminescence
but it is not able to induce Eu3+ luminescence.
Figure 5.39: Energy level diagram of Eu3+ and Yb3+ ions with the 5 D0 →7 FJ Eu3+ emissions and
the proposed Yb3+ -Eu3+ energy transfer mechanisms.
Figure 5.40: RT UC luminescence spectra of GGG: 2%Eu3+ , 1%Er3+ (a) and GGG: 2%Eu3+ , 5%Yb3+
(b) exciting at 10250 cm−1 .
Therefore, any relevant role of Er3+ in the Yb3+ to Eu3+ UC luminescence is ex-
cluded, although the efficient Yb3+ to Er3+ energy transfer compared to Yb3+ to
Eu3+ one is pointed out.
The time evolution of the 5 D4 →7 FJ Tb3+ and 5 D0 →7 FJ Eu3+ emission in GGG

and YAG singly-doped and co-doped nanoparticles has been recorded after direct
excitation at around 20365 and 16920 cm−1 for Tb3+ and Eu3+ , respectively. GGG
experimental decays are shown in Fig. 5.41.
Figure 5.41: Temporal evolution of 5 D4 → 7 FJ Tb3+ RT luminescence in GGG: 2%Tb3+ , 5%Yb3+

upon excitation at 20365 cm−1 (a), and 5 D0 → 7 FJ Eu3+ RT emission in GGG: 2%Eu3+ , 5%Yb3+ upon
excitation at 16920 cm−1 (b).
Table 5.6: Lifetimes of the 5 D4 Tb3+ and 5 D0 Eu3+ levels in GGG and YAG samples
measured at the emission energy, Eem , after direct excitation into the emitting levels.
Eem /cm−1 Sample τ /ms Sample τ /ms

18396 GGG: 2%Tb3+ , 5%Yb3+ 4.2 ± 0.2 GGG: 2%Tb3+ 4.1 ± 0.2
18396 YAG: 2%Tb3+ , 5%Yb3+ 4.9 ± 0.2 YAG: 2%Tb3+ 4.7 ± 0.2
14134 GGG: 2%Eu3+ , 5%Yb3+ 4.0 ± 0.2 GGG: 2%Eu3+ 4.1 ± 0.2
14090 YAG: 2%Eu3+ , 5%Yb3+ 4.6 ± 0.2 YAG: 2%Eu3+ 4.6 ± 0.2
Experimental data have been fitted to a single exponential for all samples and the
corresponding lifetimes are shown in Table 5.6.
The Tb3+ and Eu3+ lifetimes obtained in this work are longer than the average
values measured in doped Y3 GaO6 [39] or Li2 SiO3 [40] nanoparticles. It is also
worth mentioning that the lifetimes for Tb3+ and Eu3+ ions in nanocrystalline GGG
and YAG are significantly longer than the usual values found for the doped bulk
oxide counterparts [41]. This behavior was also observed for nanocrystalline Eu3+
doped Y2 O3 [42] and ZrO2 [43] powders and it is compatible with the small particle
size of the nanopowders, around 40 nm. In fact, this lengthening of the lifetimes
can be ascribed to a lower refractive index (neff ) surrounding the lanthanide ions
in the nanocrystalline material, compared to the micrometer size host, according
to a higher surface-to-volume ratio [42]. The lifetimes obtained by fitting the 2 F5/2
→ 2 F7/2 Yb3+ luminescence decay to a single exponential in co-doped GGG and
Table 5.7: Lifetimes of the 2 F5/2 Yb3+ excited state in GGG and YAG samples measured
detecting at 9750 cm−1 upon excitation at 10250 cm−1 .
Sample τ /µs
GGG: 2%Tb3+ , 5%Yb3+ 595 ± 10
YAG: 2%Tb3+ , 5%Yb3+ 485 ± 10
GGG: 2%Eu3+ , 5%Yb3+ 555 ± 10
YAG: 2%Eu3+ , 5%Yb3+ 480 ± 10
YAG samples are shown in Table 5.7. The same lifetime values are observed for
singly Tb3+ (or Eu3+ ) doped samples and Tb3+ -Yb3+ (or Eu3+ -Yb3+ ) co-doped
ones. Hence, there is no evidence of back-transfer from Tb3+ or Eu3+ to Yb3+ .
Tb3+ and Eu3+ UC luminescence has been measured in GGG and YAG upon exci-
tation at 10250 cm−1 in the 2 F5/2 Yb3+ level. Tb3+ and Eu3+ ions have no excited
states resonant with the 2 F7/2 → 2 F5/2 Yb3+ transition in the IR region. Therefore,
the UC emission cannot be explained by a direct transfer (GSA/ETU) from the
excited Yb3+ to a Tb3+ or Eu3+ ion. A cooperative GSA/ESA mechanism in Tb3+ -
Yb3+ dimers coupled by exchange was demonstrated to be responsible for Tb3+ UC
emission after Yb3+ excitation in Cs3 Tb2 Br9 : Yb3+ at low temperatures [12]. How-
ever, GGG and YAG crystal structures are not able to accommodate Tb3+ -Yb3+ or
Eu3+ -Yb3+ dimers. Hence, GSA/ESA in dimers can be ruled out as a mechanism
responsible for the UC luminescence. Additionally, UC phenomena in YF3 : Tb3+ ,
Yb3+ were explained as a cooperative sensitization [10]. Strek et al. proposed two
mechanisms to explain the anti-Stokes Eu3+ emission in KEu0.2 Yb0.8 (WO4 )2 upon
Yb3+ excitation, which were cooperative sensitization and a process based on Yb3+
pairs [13].
The unambiguous manner to investigate the mechanism involved in the UC lu-

minescence is to study its temporal evolution under pulsed excitation [12]. The
cooperative sensitization, a three ions process, involves energy transfer from two
sensitizers (Yb3+ ions) to the activator (Tb3+ or Eu3+ ). This is a slow process
which is active during the Yb3+ lifetime. The temporal evolution of the 5 D4 → 7 FJ
Figure 5.42: Temporal evolution of Tb3+ UC emission intensity in YAG: 2%Tb3+ , 5%Yb3+ detecting
at 18396 cm−1 after pulsed excitation at 10250 cm−1 .
green UC luminescence of Tb3+ obtained upon Yb3+ excitation at 10250 cm−1 for
the YAG: 2%Tb3+ , 5%Yb3+ sample is shown in Fig. 5.42. The rise of the emis-
sion intensity after IR excitation is a clear fingerprint of an energy transfer process.
The cooperative sensitization is proposed as the mechanism responsible for Tb3+
UC luminescence in GGG and YAG systems co-doped with Yb3+ . This mecha-
nism involves three ions and is depicted in Fig. 5.34. The cooperative sensitization
responsible for the Tb3+ UC luminescence is as follows: two Yb3+ ions in the ex-
cited state transfer non-radiatively the energy to the 5 D4 states of Tb3+ . The same
cooperative sensitization mechanism would be expected in the Eu3+ -Yb3+ UC lu-
minescence (see Fig. 5.39). The Tb3+ lifetime observed in the UC decay transients,
1.8 ± 0.1 ms, is shorter than the one obtained after direct excitation, 4.9 ± 0.2 ms.
(See Fig. 5.42 and Table 5.6). This difference is related to the fact that not all Tb3+
have necessarily an Yb3+ ion as a close neighbor. In UC measurements, only those
Tb3+ which are close to Yb3+ ions are excited, whereas after direct excitation, all
Tb3+ are involved in the luminescence [44].
The temporal evolution in Tb3+ -Yb3+ systems can be simulated by considering a

cooperative sensitization mechanism. Taking into account the four level system
schematically shown in Fig. 5.43, we can write down the four coupled differential
rate equations (eq. 5.1) describing the populations of each level, Ni :
Figure 5.43: Diagram of the cooperative sensitization of Tb3+ UC luminescence. In this process two
excited Yb3+ ions transfer non-radiatively their energy to Tb3+ ions.
dN0 N1
= −GN0 + + 2kCS N12 N2
dt τ1
dN1 N1
= GN0 − − 2kCS N12 N2
dt τ1
dN2 N3
= − kCS N12 N2 (5.1)
dt τ3
dN3 N3
= − + kCS N12 N2
dt τ3
where τ1 and τ3 represent the Yb3+ 2 F5/2 and Tb3+ 5 D4 experimental lifetimes, re-
spectively, G is the power dependence GSA rate constant and kCS is the cooperative
sensitization rate parameter. The physical meaning of these equations is the follow-
ing; the rise is related with half of the Yb3+ 2 F5/2 lifetime, that would correspond to
the lifetime of Yb3+ pairs, whereas the decay is related to the Tb3+ 5 D4 lifetime. N0
and N2 are known from Yb3+ and Tb3+ concentration values, respectively, whereas
N1 and N3 are an initial hypothesis since they cannot be experimentally measured
in an accurate manner. Values of 485 ± 10 µs (see Table 5.7) and 1.8 ± 0.1 ms
(see Fig. 5.42) have been obtained from independent measurements for τ1 and τ3 ,
respectively, in YAG: 2%Tb3+ , 5%Yb3+ . It is possible to obtain the only unknown
parameter, kCS , by fitting the time dependent evolution of the UC luminescence,
obtained experimentally after short-pulsed excitation into the Yb3+ 2 F5/2 level, to
this set of equations (Fig. 5.42). In this way, a value of 1350 s−1 has been determined
for the energy transfer rate constant, kCS N1 N2 .
112 5.3. Rare-earth and transition-metal ions co-doped systems
Conclusions
GGG and YAG nanoparticles co-doped with Tb3+ or Eu3+ and Yb3+ have been prepared
by the sol-gel Pechini’s method. The average size of the particles is around 30-40 nm, as
shown by XRD and TEM measurements. Green and blue Tb3+ emission as well as red
Eu3+ luminescence have been studied and identified. Excitation in the NIR region around
10250 cm−1 leads to strong Tb3+ and Eu3+ visible UC emission at RT. The experimental
results and theoretical rate equations model confirm the cooperative sensitization as the
UC mechanism responsible for the UC luminescence. The presence of Er3+ impurities and
its consequent luminescence upon 10250 cm−1 excitation in Eu3+ -Yb3+ co-doped systems
is not relevant for the Eu3+ UC emission.
5.3 Rare-earth and transition-metal ions co-doped

systems
In the previous section, different UC systems involving combinations of RE ions have been
studied. Opposite to f−f transitions in RE ions, d−d transitions of TM ions are much
more sensitive to the local environment, and therefore, associated UC properties might be
tuned by changing the crystal composition [45] or by applying hydrostatic pressure [46].
For instance, depending on the crystal-field strength, Mn2+ can emit from the blue-green
to the deep red spectral region. In this sense, the combination of lanthanide and TM ions
extends the UC luminescent tuning capability through changes in both energy resonances
and emitting state energies.
The first demonstration of UC luminescence in TM-RE mixed systems was observed
in Yb3+ -doped CsMnCl3 and RbMnCl3 , and in Cr3+ -Yb3+ co-doped Y3 Ga5 O12 at low
temperatures [47], [48]. A GSA/ESA process between Mn2+ -Yb3+ dimer states was es-
tablished as the main mechanism involved in the red UC luminescence in Yb3+ -doped
RbMnCl3 single-crystal [49]. This host incorporates trivalent impurities as Mn2+ -Yb3+
pairs with charge compensation vacancies. The use of manganese hosts ensures cluster-
ing since Yb3+ ions has always a Mn2+ ion as a near neighbor. The same model was
proposed to explain both Mn2+ and Yb3+ -pairs UC emission in Yb3+ -doped CsMnBr3
single crystals [50]. Evidence of GSA/ETU between Mn2+ -Yb3+ dimers was found in
Yb3+ -doped CsMnCl3 [19]. All these compounds showed UC luminescence only at low
temperature. It is well known that concentrated TM compounds exhibit luminescence
quenching as temperature increases. This phenomenon was explained by thermally acti-
vated excitation migration and subsequent transfer to non-radiative traps [51]. Therefore,
the temperature-induced luminescence quenching is inherent to concentrated manganese
systems, and occurs independently of the excitation way, either in the NIR or directly in
the Mn2+ ions in the UV-visible region, although the quenching temperature is lower in
the former case. Since the Mn2+ -Yb3+ UC systems studied up to now were only efficient
at cryogenic temperatures, the combination of TM and RE ions has not been exploited
for applications as UC luminescent materials.
Another difficulty related to the UC luminescence investigation involves finding suit-
able host lattices able to accommodate both Mn2+ and Yb3+ at well defined sites without
charge compensation. Regarding this subject, micropowders of LMA co-doped with dif-
ferent Mn2+ and Yb3+ concentrations have been prepared in this work. As we will see,
this host meets the impurities incorporation requirements. Since the as obtained powders
can be milled in the planetary ball mill subsequently, this might be a good initial step
to obtain LMA: Mn2+ , Yb3+ nanoparticles. The experimental results shown in this sec-
tion allow to establish the UC mechanisms involved in Mn2+ -Yb3+ co-doped LMA, and
determine the effect of doping concentrations on the UC efficiency.
5.3.1 Mn2+ , Yb3+ co-doped LaMgAl11 O19

LMA crystallizes in the PbFe12 O19 magnetoplumbite-type structure (hexagonal, P 63 /mmc

space group) [52]. It consists of spinel blocks separated by an intermediate layer containing
three oxygen ions, one lanthanum and one aluminum per spinel block. Mg2+ ions are
accommodated into the spinel blocks. LMA doped with Mn2+ is a well-known green
phosphor which can be excited by vacuum UV radiation and used in plasma display
panels [53]. In Mn2+ -Yb3+ co-doped LMA, Mn2+ ions occupy the tetrahedral Mg2+ sites
whereas Yb3+ replaces La3+ .
Figure 5.44: XRD patterns of LMA: 2%Mn2+ , 5%Yb3+ (a) and LMnA: 1%Yb3+ (b). Colored dots
show the experimental data and black lines the Rietveld refinements.
LMA microcrystalline samples doped with nominal concentrations of 1%Mn2+ -1%Yb3+ ,

2%Mn2+ -5%Yb3+ , and LaMnAl11 O19 (LMnA) doped with 1%Yb3+ have been prepared
by the precipitation method described in Section 3.6. LMA: 1%Mn2+ , 1%Yb3+ has also
been synthesized by a combustion route. XRD patterns of 2%Mn2+ -5%Yb3+ co-doped
LMA and 1%Yb3+ -doped LMnA are shown in Fig. 5.44. Both patterns indicate a fairly
good crystallization of synthesized powders in the magnetoplumbite structure [54]. For
the structural refinement, the ”vacancy model” published by Iyi et al. on a lanthanum
hexa-aluminate compound has been exploited [55].
The Rietveld refinement of the LMA: 2%Mn2+ , 5%Yb3+ XRD pattern is consis-
tent with the magnetoplumbite structure with lattice parameters a=5.5602(5) Å and
c=22.055(4) Å, but also reveals small traces ( ∼4% volume fraction) of Al2 O3 (R−3c space
group). Similarly, the XRD pattern of LMA: 1%Mn2+ , 1%Yb3+ (data not shown) reveals
the magnetoplumbite structure with lattice parameters a=5.5678(2) Å and c =22.026(2)
Å of the as obtained powders, and traces of Al2 O3 are also detected. The divalent ion
occupies exclusively the 4f site (0.5 occupancy factor) together with an Al3+ ion and both
ions are tetrahedrally coordinated by oxygens with a C3v point symmetry [52].
For the LMnA powders containing Mn2+ instead of Mg2+ similar results are obtained;
97% of LMnA magnetoplumbite structure and small traces of Al2 O3 . Mn2+ ions also
occupy only the 4f site (C3v symmetry) [56]. For this phase the refined cell parameters
are: a = 5.5818(5) and c=22.070(3) Å. The increased cell parameters observed for LMnA
in comparison with LMA reflect the bigger ionic radius of Mn2+ with respect to Mg2+ .
Figure 5.45: Unit cells for LMA (left) and LMnA (right) crystals. The tetrahedral coordination of
Mg2+ or Mn2+ by oxygens is illustrated.
The unit cells of the LMA and LMnA crystal structures are shown in fig. 5.45. There
are two different sites for the lanthanides in both LMA and LMnA phases. In the LMA
sample, the principal site (0.49 occupancy) has D3h symmetry and a 12-fold coordination
number, whereas the satellite La site (0.115 occupancy) has a lower point symmetry (C2v )
[52]. In the case of LMnA compound, the satellite La site has the same point symmetry
as in LMA, but with lower occupancy factor (0.05), whereas there is a change in the
point symmetry (from D3h to Cs ) in the principal La site. From our structural refinement
the following distances are obtained: La(1)-Mg=5.844 Å and La(2)-Mg=5.728 Å for LMA
and La(1)-Mn=5.795 Å and La(2)-Mn=5.724 Å for LMnA. The shortest La-La distance is
11.337 Å and 11.324 Å for LMA and LMnA, respectively. All these distances are relevant
for Mn2+ -Yb3+ interaction and the subsequent energy transfer process.
Optical properties
Luminescence, excitation and lifetime measurements have been carried out on different
Mn2+ -Yb3+ co-doped LMA samples. An extended spectroscopic investigation to clarify
the effects of Mn2+ and Yb3+ concentrations as well as temperature on UC luminescence

properties is presented.
Figure 5.46: RT spectra of LMA: 1%Mn2+ , 1%Yb3+ . IR excitation Yb3+ spectrum detecting Mn2+
UC emission (a) 4 T1 → 6 A1 Mn2+ UC luminescence after IR excitation at 10250 cm−1 (b). RT Mn2+
direct excitation spectrum monitoring the Mn2+ emission at 19455 cm−1 (c). Mn2+ peaks are labeled
according to the Tanabe-Sugano diagram for tetrahedral Mn2+ with ∆=5840 cm−1 and B=730 cm−1 .
Mn2+ emission spectrum after direct excitation at 424 nm (d). RT 2 F5/2 → 2 F7/2 Yb3+ luminescence
(e).
Figure 5.46(b) and (d) shows the RT Mn2+ luminescence spectra of LMA micro-
crystals prepared by precipitation and doped with 1%Mn2+ and 1%Yb3+ upon IR
and visible excitation, respectively. The green broad band emission extended from
17000 to 20500 cm−1 , is assigned to the 4 T1 → 6 A1 Mn2+ transition in tetrahedral
coordination, and the high energy peaks observed in the UC spectra (Fig. 5.46(b)),
are due to Yb3+ -Yb3+ cooperative luminescence [57], [58], [59], [60]. As far as we
know, this is the first observation of green Mn2+ UC luminescence upon Yb3+ ex-
citation at RT. Figure 5.46(a) shows the excitation spectrum in the IR region of
the Mn2+ UC luminescence detecting at 19455 cm−1 . Figure 5.46(c) shows the RT
excitation spectrum of the 4 T1 → 6 A1 Mn2+ luminescence which proves the assign-
ment of the green luminescence to a Mn2+ ion in tetrahedral coordination. Figure
5.46(e) displays the 2 F5/2 → 2 F7/2 Yb3+ emission of the LMA: 1%Mn2+ , 1%Yb3+
upon excitation at 10250 cm−1 .
Figure 5.47: RT luminescence spectra of LMA: 1%Mn2+ , 1%Yb3+ prepared by precipitation (black)
and combustion (red) upon excitation at 23580 cm−1 (a) and 10250 cm−1 (b). * is an artifact and
represents the laser at twice the excitation frequency. Measurements were carried out under the same
experimental conditions.
Since high purity green UC luminescence from tetrahedrally coordinated Mn2+ ions
has been demonstrated in LMA: 1%Mn2+ , 1%Yb3+ upon Yb3+ excitation, we have
performed different attempts to increase the UC efficiency by changing the synthesis
method or the dopant concentration. Figure 5.47 compares the RT luminescence
spectra of LMA: 1%Mn2+ , 1%Yb3+ micropowders prepared by precipitation and
combustion upon visible and IR excitation. It can be observed that for both excita-
tion energies the sample prepared by precipitation has higher luminescence intensity
than the one by combustion. Moreover, UC emission intensity is about an order of
magnitude higher for precipitation powders. Figure 5.48 compares the RT lumines-
cence spectra of co-doped LMA prepared by precipitation for different Mn2+ and
Yb3+ concentrations upon visible and IR excitation. No shift in energy or change
in bandwidth is observed comparing the Mn2+ luminescence after direct and UC
Figure 5.48: RT luminescence of Mn2+ -Yb3+ co-doped LMA upon excitation at 23580 cm−1 (a) and
10250 cm−1 (b) for the following concentrations: 1%Mn2+ -1%Yb3+ (black), 2%Mn2+ -5%Yb3+ (red) and
99%Mn2+ -1%Yb3+ (blue). * is a laser artefact appearing at twice the excitation frequency. Measurements
were carried out under the same experimental conditions.
excitation. Considering the Mn2+ sensitivity to surrounding, this indicates that

both direct and UC luminescence are originated from Mn2+ ions having the same
environment, what is an indication of the homogenous distribution of Mn2+ and
Yb3+ impurities in LMA. The green Mn2+ emission in LMA: 2%Mn2+ , 5%Yb2+ is
centered at 19430 cm−1 while the same band appears about 50 cm−1 towards higher
energy for LMnA: 1%Yb3+ according to expectations due to the smaller ionic radius
of Mg2+ compared to Mn2+ .
It is worth noting that Mn2+ emission is obtained for all doping concentrations upon
excitation in both 23580 cm−1 and 10250 cm−1 bands even above RT. For direct
Mn2+ excitation (Fig. 5.48(a)), the RT luminescence intensity ratio of the pure
manganese compound (LMnA) to the Mn2+ -doped LMA has been found to be ap-
proximately 1:5, while the intensity for 1%Mn2+ and 2%Mn2+ compounds is roughly
the same. This drastic reduction of emission intensity could be initially ascribed
to thermally activated energy migration and partial trapping into non-luminescent
impurities as commonly occurs in concentrated manganese systems. Figure 5.48(b)

shows that UC intensity in 2%Mn2+ -5%Yb3+ co-doped LMA is about an order
of magnitude higher than the UC in 1%Mn2+ -1%Yb3+ co-doped LMA. Consider-
ing crystalline quality and emission efficiency, LMA co-doped with 2%Mn2+ and
5%Yb3+ doping level and prepared by precipitation provides an optimum efficient
UC system.
• Temporal evolution.
The temporal evolution of the RT 4 T1 → 6 A1 Mn2+ green luminescence has been

recorded after direct Mn2+ excitation into the 4 A1 , 4 E levels at 23580 cm−1 , and IR
excitation into the 2 F7/2 → 2 F5/2 Yb3+ transition at 10205 cm−1 , with 10 ns short
pulses for LMA: 1%Mn2+ , 1%Yb3+ , LMA: 2%Mn2+ , 5%Yb3+ and LMnA: 1%Yb3+
samples (Fig. 5.49). The intensity time-dependence of the Mn2+ emission after
visible excitation shows a single exponential decay with lifetimes of τ =6.2 ± 0.1 ms
for LMA: 1%Mn2+ , 1%Yb3+ and LMA: 2%Mn2+ , 5%Yb3+ , and τ =3.8 ± 0.1 ms for
LMnA: 1%Yb3+ (Fig. 5.49(a), (b) and (c)). However, the time dependence of the
UC Mn2+ luminescence after IR excitation presents a different behavior comparing
diluted and undiluted systems. A single exponential decay is observed in LMA:
1%Mn2+ , 1%Yb3+ (τ =5.8 ± 0.2 ms, Fig. 5.49(d)) and LMA: 2%Mn2+ , 5%Yb3+
(τ =5.2 ± 0.2 ms, Fig. 5.49(e)). This lifetime is significantly faster than the one
obtained by direct Mn2+ excitation (τ =6.2 ms) as it was previously observed in
Yb3+ -doped RbMnCl3 . It is probably related to an increase in the emission proba-
bility due to the Mn2+ -Yb3+ interaction, or to an additional Mn2+ → Yb3+ decay
path, within the UC-efficient clusters [49]. In the case of LMnA: 1%Yb3+ , a lu-
minescence intensity rise before decaying is clearly detected (Fig 5.49(f)). These
differences are crucial to identify the UC mechanism involved in these systems as it
will be discussed later on.
Figure 5.50 shows the temporal evolution of the Yb3+ -Yb3+ cooperative lumines-
cence in LMA: 2%Mn2+ , 5%Yb3+ upon IR modulated excitation at 10205 cm−1 at
15 K (a) and RT (b). The decay can be described by a single exponential function
Figure 5.49: RT temporal evolution of the normalized Mn2+ 4 T1 → 6 A1 emission intensity after pulsed
excitation at 23580 cm−1 in LMA: 1%Mn2+ , 1%Yb3+ (a), LMA: 2%Mn2+ , 5%Yb3+ (b) and LMnA:
1%Yb3+ (c). And temporal behavior of the Mn2+ UC emission after pulsed excitation at 10205 cm−1 in
LMA: 1%Mn2+ , 1%Yb3+ (d), LMA: 2%Mn2+ , 5%Yb3+ (e) at RT, and LMnA: 1%Yb3+ (f) at both 15
K and RT. The insets show the same data in semi-logarithmic scale. Continuous lines are the results of
different fitting procedures (see text).
Figure 5.50: Temporal behavior of the normalized Yb3+ -pairs emission in LMA: 2%Mn2+ , 5%Yb3+
after pulsed excitation at 10205 cm−1 at 15 K (a) and RT (b). The insets show the same data in
semi-logarithmic scale.
with lifetimes of τ =470 ± 10 µs at 15 K and τ =220 ± 10 µs at RT.
In LMA: 2%Mn2+ , 5%Yb3+ a lifetime of 5.2 ms has been obtained for the Mn2+ UC
emission (Fig. 5.49(e)) while the RT lifetime of Yb3+ -Yb3+ cooperative luminescence
is 220 µs (Fig. 5.50(b)). Since both lifetimes differ by an order of magnitude, we can
separate both UC luminescence by time-resolved spectroscopy. Figure 5.51 shows
both Mn2+ green emission (broad band centered at 19430 cm−1 ) and Yb3+ -pairs
luminescence (higher energy peaks above 19500 cm−1 ) upon IR excitation at 10205
cm−1 taken at different delay times after the excitation pulse. It can be seen that
the Yb3+ -Yb3+ cooperative emission intensity decreases much faster than the Mn2+
UC luminescence intensity according to their lifetimes. The figure clearly illustrates
this phenomenon showing the rapid decay of Yb3+ -pairs luminescence in a ms time
scale while the Mn2+ UC emission remains almost constant for this time.
• Temperature dependence.
The temperature dependence of luminescence intensity and lifetime presents dif-

ferent trends for co-doped LMA at low concentrations of Mn2+ and pure LMnA
microcrystalline samples. Figure 5.52 shows the temperature dependence of the
normalized emission intensity (I/Imax ) and the lifetime (τ ) of the 4 T1 → 6 A1 Mn2+
transition in LMA: 2%Mn2+ , 5%Yb3+ and LMnA: 1%Yb3+ upon IR excitation at
10205 cm−1 in the 15-600 K range. It must be noted that for LMA: 2%Mn2+ ,
Figure 5.51: 15 K UC time-resolved emission of LMA: 2%Mn2+ , 5%Yb3+ exciting at 10205 cm−1 using
different delay times after the short excitation pulse.
Figure 5.52: Temperature dependence of the normalized Mn2+ UC luminescence intensity (red circles)
and Mn2+ UC emission lifetime (open red squares) in LMA: 2%Mn2+ , 5%Yb3+ . And temperature
dependence of the normalized Mn2+ UC luminescence intensity (blue triangles) and Mn2+ UC emission
lifetime (blue squares) in LMnA: 1%Yb3+ after 10205 cm−1 excitation. The lines correspond to fittings
according to eqs. 5.2, 5.3 and 5.4 (see text).
5%Yb3+ the UC luminescence intensity remains almost constant up to RT. Al-

though above this temperature the quenching process becomes important and the
UC luminescence intensity starts to decrease, it still remains up to 600 K. The as-
sociated UC lifetime follows a similar trend; it diminishes slightly from 10 to 300 K
and steeply decreases above this temperature. Both the intensity and the lifetime
have a similar behavior pointing out the non-radiative character of the UC lumines-
cence decrease. An analogous behavior is detected for both lifetime and intensity
in LMA: 1%Mn2+ , 1%Yb3+ (data not shown for clarity). The quenching process
of the Mn2+ UC luminescence is rather different in LMnA: 1%Yb3+ as shown in
Fig. 5.52. As it usually happens in concentrated materials, the Mn2+ emission is
thermally quenched at much lower temperatures (T <100 K) than in Mn2+ diluted
samples. This is an intrinsic behavior of concentrated manganese and it is usually
accompanied by a similar lifetime decrease with temperature [51], [61]. However,
since in this case the lifetime remains pretty constant until RT, the reduction in
the intensity due to migration and subsequent quenching is unlikely. Instead, the
photoluminescence dwindle above 100 K which is induced either by direct excita-
tion into the 4 A1 , 4 E Mn2+ state or via UC in LMnA: 1%Yb3+ , must be ascribed
to a decrease in the effective Mn2+ excitation with temperature. As it is shown
below, this puzzling I(T ) behavior between 0 and 300 K can be phenomenologically
described through a thermal deactivation process for Mn2+ excitation.
In order to model Mn2+ luminescence in LMA: 2%Mn2+ , 5%Yb3+ let us assume that
the radiative lifetime, τR , does not change significantly with temperature. According
to a model proposed by Mott, thermally activated non-radiative processes can be
described by an activation energy, Ea , with a pre-exponential frequency factor, p
[62]. Consequently, the lifetime and intensity dependence on temperature have been
fitted according to Eqs. 5.2 and 5.3. In both cases the fitted parameters are Ea =
0.26 eV and p = 180×103 s−1 for LMA: 2%Mn2+ , 5%Yb3+ .
1
τ (T ) = 1 (5.2)
τR
+ p · exp(−Ea /KB T )
I(T ) 1
= (5.3)
Imax 1 + p · τR exp(−Ea /KB T )
It must be mentioned that this simple model explains why I(T ) and τ (T ) behave
with temperature in the same manner. Both the activation energy (Ea = 0.26 eV)
and the Mn2+ dilution suggest that the non-radiative process is associated with mul-
tiphonon relaxation within (MnO4 ) units. In this analysis, it is assumed that the
excited Mn2+ via either UC or direct excitation is approximately temperature inde-
pendent. Nevertheless, in the case of LMnA: 1%Yb3+ , eq. 5.3 must account for the
fraction of excited Mn2+ as a function of temperature: ER (T ) = [Mn](T )/[Mn](0)
with [Mn] being the concentration of excited Mn2+ . Hence, eq. 5.3 transforms to
I(T ) ER (T )
= (5.4)
Imax 1 + p · τR exp(−Ea /KB T )
On the assumption that the up-converted Mn2+ does not change with temperature
(ER = 1), we then obtain the same variation for I(T ) and τ (T ) as eq. 5.2 and 5.3.
However, this is not the case if ER varies with temperature. The I(T ) reduction
above 100 K observed in LMnA: 1%Yb3+ keeping τ (T ) constant up to RT, can
be explained through this model if we consider that Mn2+ excitation is thermally
deactivated: ER = (1 − α exp(−∆/KB T )) where ∆ is the activation energy and α
is the pre-exponential factor. The experimental I(T ) data (blue) behaves in this
way for values of ∆ = 5.6 meV and α = 1.2. Although the present model makes
the reduction of intensity and the lifetime constancy compatible, its microscopic
origin is unclear. This thermal behavior is unusual for Mn2+ concentrated systems
in which I(T ) correlates with τ (T ).
• UC mechanism.
The RT 4 T1 → 6 A1 Mn2+ UC luminescence intensity versus the excitation power

density at 10250 cm−1 for LMA: 2%Mn2+ , 5%Yb3+ is plotted on a double logarith-
mic scale in Fig. 5.53. This system presents a quadratic power dependence below
1.5 Wcm−2 which is the typical behavior of a two-photon excitation process in the
Figure 5.53: Excitation power density dependence of the RT Mn2+ UC emission in LMA: 2%Mn2+ ,
5%Yb3+ upon excitation with a LD at 10250 cm−1 .
Figure 5.54: Energy level diagram in a Mn2+ -Yb3+ dimer notation with the Mn2+ emission and the
proposed UC mechanisms.
low-power regime. The slope decreases from 2 to 1.5 at higher excitation power
densities as explained before [24], [25]. Since excitation occurs into 2 F5/2 Yb3+
states, and emission takes place from 4 T1 Mn2+ states, the active UC mechanism in
the three studied samples, LMA: 1%Mn2+ , 1%Yb3+ , LMA: 2%Mn2+ , 5%Yb3+ and
LMnA: 1%Yb3+ must involve the participation of both Yb3+ and Mn2+ ions. Mn2+
has no intermediate resonant states with Yb3+ ; hence, GSA/ESA and GSA/ETU
mechanisms in pure ions must be ruled out in these systems. A simple model based
on an exchange-coupled Mn2+ -Yb3+ dimer state was previously proposed to explain
the experimental behavior in Yb3+ -doped RbMnCl3 , CsMnBr3 and CsMnCl3 [47],
[49].
The UC emission in LMA: 1%Mn2+ , 1%Yb3+ and LMA: 2%Mn2+ , 5%Yb3+ (Fig.
5.49 (d) and (e)) shows an immediate decay after the laser pulse with no rise due
to energy transfer longer than 80 ns (low detection limit). According to the LMA
crystal structure (Fig. 5.45), the shortest Mn2+ -Yb3+ distance for Mn2+ -O-O’-Yb3+
dimers in Mn2+ -Yb3+ co-doped LMA is 5.728 Å. In agreement with Luzón et al.
and Hehlen and co-workers, superexchange distances can be as long as 6-8 Å; thus,
dimers in co-doped LMA are able to exhibit a superexchange pathway [63], [64].
Taking into consideration a GSA/ESA-type mechanism in Mn2+ -Yb3+ dimers, the
GSA step is predominantly a single ion process in Yb3+ , but the ESA step requires
the mixing of Mn2+ and Yb3+ states. The rise of the emission intensity after the
excitation pulse in LMnA: 1%Yb3+ (Fig. 5.49(f)) evidences the contribution of
some non-radiative energy transfer process after the pulse to the UC mechanism.
The involved processes can be GSA/ETU in a dimer or cooperative sensitization.
Once more, the Mn2+ -Yb3+ distance in LMnA (5.724 Å) is short enough to allow
dimer formation, hence we propose the GSA/ETU as the actual mechanism. The
Mn2+ lifetime is shorter in LMnA: 1%Yb3+ than in 1% or 2% Mn2+ -doped LMA;
this evidences a larger exchange contribution for the pure manganese system. The
fact that the intensity rise does not start from zero is noteworthy. This implies
that there must be an additional excitation within the duration of the laser pulse
governed by a GSA/ESA mechanism and it is another evidence of dimer formation.
Considering that the lifetime of the 2 F7/2 − 4 T1 dimer state (see Fig. 5.54) is
approximately the same as Mn2+ lifetime after direct excitation, it is possible to
estimate the GSA/ESA and GSA/ETU contributions to the total intensity as we
have done for Y2 O3 : Er3+ , Yb3+ nanoparticles (Section 5.2.1) [19]. At 15 K, about
23% of the total UC luminescence is due to excitation within the pulse, GSA/ESA,
while contributions of 33% and 67% have been obtained at RT for GSA/ESA and
GSA/ETU mechanisms respectively.
Figure 5.54 shows a schematic representation of an exchange-coupled Mn2+ -Yb3+

dimer and an Yb3+ nearby monomer along with the proposed mechanisms for the UC
luminescence; GSA/ESA for LMA: 1%Mn2+ , 1%Yb3+ and LMA: 2%Mn2+ , 5%Yb3+
and both GSA/ESA and GSA/ETU for LMnA: 1%Yb3+ . The dimer intermediate
state are dominantly localized on Yb3+ while the higher excited states are mainly
localized on Mn2+ ions [49].
Figure 5.55: Relevant levels involved in the GSA/ESA and GSA/ETU UC processes in LMnA: 1%Yb3+ .
The temporal evolution of the LMnA: 1%Yb3+ can be simulated considering GSA/ESA
and GSA/ETU mechanisms according to the five level system shown in Fig. 5.55
involving an Yb3+ ion and a Mn2+ -Yb3+ dimer. Within this model, the coupled
differential rate equations describing the population of each level, Ni can be written
as:
dN0 N1
= −GN0 + + WETU N1 N3
dt τ1
dN1 N1
= GN0 − − WETU N1 N3
dt τ1
dN2 N3 N4
= −GN2 + + (5.5)
dt τ3 τ4
dN3 N3 N4
= GN2 − − EN3 + − WETU N1 N3
dt τ3 τ43
dN4 N4 N4
= EN3 − + WETU N1 N3 −
dt τ43 τ4
where τi represents the lifetime of each level Ni , G and E are the power dependent
GSA and ESA rate constants, respectively, and WETU is the two-center energy-
transfer process parameter. τ1 and τ3 are known from the experimentally measured
Yb3+ lifetime; 940 µs at 15 K and 370 µs at RT, which are significantly longer than
those obtained for LMA: 2%Mn2+ , 5%Yb3+ , while τ4 is the Mn2+ emission lifetime;
4.4 ms at 15 K and 3.8 ms at RT. By fitting the time dependent evolution of the
UC luminescence in LMnA: 1%Yb3+ to this set of equations, we obtain an energy
transfer rate, WETU N1 N3 , of 2310 s−1 and 3125 s−1 for RT and 15 K respectively
(Fig. 5.49(f)).
Nevertheless, the UC intensity evolution depicted in Fig. 5.49(f) can be expressed

in terms of Mn2+ and Yb3+ lifetimes. In fact, the rate equation for N4 (t ≥ 0) can
be solved analytically under the assumption that Yb3+ populations, N1 and N3 ,
are similar (N1 ≈ N3 ), both decaying as N3 = N30 exp(−t/τ3 ). This approxima-
tion yields the following time-dependent Mn2+ population: N4 = N40 (exp(−t/τ ) −
1 1 1
α exp(−2t/τ3 )). The parameter τ
= τ4
+ τ43
, governing the temporal decay of Fig.
5.49(f), is directly related to the Mn2+ lifetime, whereas τ3 controls the intensity
rise after t = 0. The parameters α and N40 determines the Mn2+ population at t = 0
due to GSA/ESA. By fitting the intensity temporal dependence in Fig. 5.49(f) to
this approximate equation, the following parameters are obtained: τ = 4.2 ms and
τ3 = 960 µs at RT, and τ = 4.7 ms and τ3 = 435 µs at 15 K. These values are quite
similar to Mn2+ and Yb3+ lifetimes, respectively, what supports the proposed UC
GSA/ETU mechanism in LMnA: 1%Yb3+ .
Conclusions
The spectroscopic study carried out on LMA microcrystals doped with different Mn2+
and Yb3+ concentrations indicates that Mn2+ UC luminescence is very efficient in these
systems even at RT. LMA crystallizes with the magnetoplumbite crystal structure where
Mn2+ ions are tetrahedrally coordinated by oxygen atoms. Mn2+ UC luminescence and
Yb3+ -pairs emission have been detected in all samples. It has been demonstrated that
UC luminescence takes place via GSA/ESA mechanism for LMA: 1%Mn2+ , 1%Yb3+ and
LMA: 2%Mn2+ , 5%Yb3+ , whereas an additional GSA/ETU contribution participates in
LMnA: 1%Yb3+ . This conclusion is based on time-resolved spectroscopy and impurity
concentration dependent experiments. The theoretical rate equations model confirm the
different proposed UC mechanisms involving Mn2+ -Yb3+ dimer formation. In all LMA
and LMnA systems the luminescence via UC is observed up to RT and above 500 K
in the case of LMA: 2%Mn2+ , 5%Yb3+ , being this UC emission an order of magnitude
stronger than any other. Mn2+ UC luminescence in LMA: 2%Mn2+ , 5%Yb3+ remains
very intense up to RT but at higher temperatures both the emission intensity and lifetime
decrease in a similar manner, pointing out the non-radiative character of the quenching.
On the contrary, an unusual decrease in the Mn2+ luminescence above 100 K with a
constant lifetime has been detected in LMnA: 1%Yb3+ . This cannot be ascribed to
Mn2+ migration and subsequent quenching, and it must be related to a decrease of Mn2+
pumping efficiency.
5.4 Transition-metal ions doped nanoparticles

Optical properties of materials containing TM ions depend on different factors such as site
symmetry and crystal-field strength. The application of high pressure to these systems
changes the crystal-field strength experienced by the TM ion, and it may also modify
the symmetry. Hence, it is a powerful tool to obtain information about TM luminescent
properties by continuously changing bond distances as well as crystal-field strengths or
covalency [65].
Depending on the crystal-field strength, the lowest Cr3+ emitting level can be the 4 T2
or the 2 E state (see Fig. 2.5) [51]. Each excited state exhibits different luminescence
properties due to the different electron-lattice coupling. Typical emission features for the
weak crystal-field case are a fast decay and broad emission band due to the spin-allowed
4
T2 → 4 A2 transition. In contrast, the narrow R-lines emission associated with the spin-
forbidden transition from the 2 E level to the 4 A2 ground state is expected for the strong
crystal-field scenario. An interesting situation arises for intermediate crystal-field, when
the crystal-field strength on Cr3+ ions is near the 4 T2 − 2 E excited-state crossover (ESCO),
and the energy separation between both states is small. In this case, the 4 T2 state may
interact with the 2 E state through thermal population effects [66] or through spin-orbit
coupling [67], and the lowest emitting state is a mixture of both excited levels [65]. The
effect of the 2 T1 state has been neglected in this work for simplicity. Oxide garnets are
attractive host materials to investigate ESCO processes since they provide intermediate
130 5.4. Transition-metal ions doped nanoparticles
crystal-fields around Cr3+ which are close to the 4 T2 -2 E crossing point, resulting in a
strong coupling of the excited states [68]. Concretely, in bulk GGG: Cr3+ , the 4 T2 state
has been seen to be located approximately 300-400 cm−1 above the 2 E level (see Fig. 2.5)
[69], [70].
The luminescence spectra associated with the 2 E → 4 A2 transition of octahedral Cr3+
in YAG: Cr3+ single crystal was investigated by Wall et al. as a function of temperature
[66]. The spectra presented sharp R-lines 2 E → 4 A2 emission, as well as low energy
sidebands. These sidebands were more important at higher temperatures and their nature
was discussed since it could be attributed to both vibronic transitions and exchange
coupled Cr3+ pairs. These authors also pointed out another possibility; the emission from
the 4 T2 state. However, only vibronic processes were established to be responsible for the
sidebands in this case.
As it was mentioned before, the application of high pressure allows to continuously
tune the Cr3+ crystal-field strength, and hence to investigate the 2 E − 4 T2 coupling.
Cr3+ ESCO was first observed by Dolan et al. in K2 NaGaF6 : Cr3+ at 6.1 GPa and 154
K [71]. Moreover, a complete change from broadband to ruby-like emission was observed
in LiCaAlF6 : Cr3+ at RT and 28 GPa [67]. The RT high pressure luminescence study in
bulk GGG: Cr3+ was carried out by Hömmerich et al. [69]. The emission spectra showed a
broad band at ambient pressure, indicative of the 4 T2 → 4 A2 transition, and the reduction
of this broad band intensity accompanied by the increase of R-lines emission, 2 E → 4 A2 ,
induced by the application of pressure up to 10 GPa. This change from broad band
emission to ruby-like emission was correlated with the increase in the emission lifetime
with pressure, corresponding to the transformation from the spin-allowed 4 T2 → 4 A2 to
the spin-forbidden 2 E → 4 A2 transition. As far as we know, no studies on GGG: Cr3+
nanoparticles under high pressure have been carried out so far.
5.4.1 Cr3+ -doped Gd3 Ga5 O12

0.5%Cr3+ -doped GGG nanocrystals have been prepared by Pechini’s method as described
in Section 3.4. It was stated that for Nd3+ -Cr3+ co-doped GGG single crystals, Nd3+
replaces Gd3+ sites while TM3+ , in particular Cr3+ , occupies Ga3+ sites [30]. Hence,
unlike RE3+ doped GGG, the amount of Cr3+ ions (0.5% mol) is with respect to Ga3+ .
From XRD patterns, the nanoparticles has been seen to exhibit the garnet structure with
an average size of 30 nm, similarly to the case of lanthanides doped GGG (see Fig. 5.27).
Optical properties
Since Heer et al. detected UC emission at low temperatures in Cr3+ -Yb3+ co-doped
Y3 Ga5 O12 and YAG single crystals [48], [72], we initially focused on the UC luminescence
study in GGG: Cr3+ , Yb3+ nanoparticles. However, no UC emission was detected in these
samples even at low temperatures. Owing to the fact that the crystal-field strength on
the Cr3+ site is very close to the ESCO, we have investigated in this work the optical
properties of Cr3+ -doped GGG nanoparticles as a function of temperature and pressure
in order to establish the effect of the ESCO on the optical behavior of nanocrystalline
GGG. Luminescence and lifetime measurements have been performed on GGG: 0.5%Cr3+
nanoparticles in the 25-300 K range as well as under hydrostatic pressure up to 20 GPa
at RT.
Figure 5.56: RT excitation spectrum of GGG: 0.5%Cr3+ detecting Cr3+ emission at 14085 cm−1 . The
arrow corresponds to the R-lines emission energy.
The RT excitation spectrum of GGG: 0.5%Cr3+ nanoparticles monitoring lumines-

cence at 14085 cm−1 is shown in Fig. 5.56. The two broad bands centered around
15800 and 21700 cm−1 correspond to the spin allowed 4 A2 → 4 T2 and 4 A2 → 4 T1 tran-
sitions respectively, in an octahedral environment, with ∆ = 15800 cm−1 , B = 705
cm−1 and C = 2880 cm−1 . The ∆
B
= 22.4 value is slightly higher than the ESCO ratio,
∆
B
(ESCO) = 22 (Fig. 2.5), close to the situation observed in the bulk.
Figure 5.57: Temperature dependence of the GGG: 0.5%Cr3+ normalized luminescence spectra upon
excitation at 21980 cm−1 at AP.
• Temperature dependence.
Figure 5.57 shows the temperature dependence of the luminescence spectra of GGG:
0.5%Cr3+ nanoparticles upon excitation at 21980 cm−1 in the 25-300 K range. At
RT, only luminescence from the 4 T2 state is detected in other Cr3+ -doped garnets
such as Gd3 Sc2 Ga3 O12 or La3 Lu2 Ga3 O12 [69], [73], or fluorides [67], according to the
weaker crystal-field strength. In GGG on the contrary, since the 4 T2 state is situated
above the 2 E one, besides the broadband emission due to the 4 T2 → 4 A2 Cr3+
transition, sharp lines corresponding to the 2 E → 4 A2 emission are also observed.
However, when temperature decreases, the intensity of the broad luminescence from
4
T2 is reduced, and its diminution is accompanied by the evolution of structured
narrow R-lines emission. The pure 4 T2 emission almost disappears at 100 K and the
Table 5.8: Lifetimes of the GGG: 0.5%Cr3+ nanocrystals corresponding to the R-line
and sidebands emission in the 25-300 K temperature range. The data for the bulk are
extracted from Di Bartolo’s presentation at the Spectroscopy Workshop (Erice 2010) [74].
T/K τ /ms R-line τ /ms R-line τ /ms sidebands τ /ms sidebands

−1 −1
(14410 cm ) bulk (13985 cm ) bulk
25 1.68 ± 0.03 1.35 1.69 ± 0.04 1.20
50 1.65 ± 0.04 1.30 1.54 ± 0.03 1.15
100 1.21 ± 0.02 0.85 1.13 ± 0.04 0.80
150 0.59 ± 0.02 0.40 0.57 ± 0.02 0.40
200 0.34 ± 0.02 0.27 0.33 ± 0.03 0.27
250 0.27 ± 0.02 0.19 0.21 ± 0.03 0.19
300 0.19 ± 0.01 0.14 0.19 ± 0.03 0.14
spectra below that temperature show sharp R-line emission with the associated low
energy bands. As it was discussed by Wall et al., the sidebands in the Cr3+ spectra
could be attributed to vibronic transitions, exchange coupled Cr3+ pairs, or emission
from the 4 T2 state [66]. Cr3+ pairs emission can be ruled out in our case because of
the low Cr3+ concentration, and the fact that some peaks appears up to 750 cm−1 .
Surprisingly, the thermal behavior of Cr3+ luminescence is completely analogous
to the Cr3+ emission spectra upon increasing pressure observed by Hömmerich et
al. for bulk Cr3+ -doped GGG. In that work, the 4 T2 emission transformed into 2 E
emission above 10 GPa (Fig. 2 in Ref. [69]).
The temporal evolution of the ruby-like Cr3+ luminescence (14410 cm−1 ) as well as
the emission from the sidebands (13985 cm−1 ), after pulsed excitation with the OPO
at 21980 cm−1 has been recorded in the 25-300 K temperature range. Experimental
data have been fitted to a single exponential, I(t) = I0 exp( −t
τ
) and the best fits
are shown in Table 5.8. The experimental lifetimes for nanocrystals have been
compared with those reported for GGG: Cr3+ in bulk. The values in the case of the
nanocrystals are significantly longer than those of the bulk (unknown concentration)
[74].
It is important to point out that very similar values have been measured for the life-
times of both emissions; R-lines and low energy sidebands. This is clearly observed
Figure 5.58: 25 K time-resolved luminescence of GGG: 0.5%Cr3+ upon pulsed excitation at 21980
cm−1 using different delay times after the excitation pulse. (Experimental details in Section 4.3.3).
in the 25 K time-resolved luminescence spectra depicted in Fig. 5.58 which shows

both R-line luminescence at around 14410 cm−1 , and sidebands emission centered
at 19500 cm−1 , upon excitation at 21980 cm−1 , obtained at different delay times
after the excitation pulse. On the basis of this temporal behavior, the existence of
different Cr3+ sites as well as Cr3+ pairs luminescence can be rejected.
Figure 5.59 shows the temperature dependence of the Cr3+ luminescence lifetime
in GGG: Cr3+ nanocrystals. The same lifetime is expected for the emission from
the R-line region and from the sidebands if the 4 T2 state is thermally populated,
or if both 2 E and 4 T2 states are coupled through spin-orbit interaction [66], [69].
However, the thermal coupling model cannot account for changes in lifetime when
one of the states is completely depopulated, which is the case of low temperature
data [65]. A model including spin-orbit coupling of the excited states leads to a
common lifetime dominated by the magnitude of the zero phonon line (ZPL) energy
separation between the 2 E and 4 T2 states, ∆E = EZPL (4 T2 )−EZPL (2 E) = 490 cm−1 .
This value is a little higher than the previously reported for the bulk, 300-400 cm−1
[69], [70]. The mixed states resulting from the spin-orbit interaction are given by:
Figure 5.59: Temperature dependence of the Cr3+ emission lifetime in GGG: Cr3+ nanoparticles. The
line corresponds to the least-square fitting to eq. 5.9 (see text).
Ψ1 = c|2 Ei + d|4 T2 i (5.6)
Ψ2 = d|2 Ei − c|4 T2 i (5.7)
being the spin-orbit mixing coefficients, c and d:
r 1
1 ∆E 2
c= 1+ (5.8)
2 ∆ESO
r 1
1 ∆E 2
d= 1−
2 ∆ESO
p
2
with ∆ESO = ∆E 2 + 4VSO , and VSO being the spin-orbit interaction energy.
Considering the lifetimes of each state, τ1 and τ2 , the common lifetime can be
described as [51], [65]:
1
+ 3 τ12 exp −∆E SO

1 τ1 kT
= (5.9)
1 + 3 exp −∆E

τ kT
SO
with
1 1 1
= c2 + d 2 (5.10)
τ1 τE τT
1 1 1
= d 2 + c2
τ2 τE τT
where τE and τT are the intrinsic decays of the 2 E and 4 T2 levels, respectively.
The obtained fitting parameters are VSO = 58 cm−1 , τ1 = 1.67 ms and τ2 = 80 µs.
The spin-orbit interaction is reduced in comparison with the typical free ion value,
215 cm−1 , as a consequence of the dynamic Jahn-Teller effect, which is known as
the Ham effect [75], [76]. Since the 4 T2 → 4 A2 transition is spin-allowed with a short
lifetime, whereas the 2 E → 4 A2 one is spin-forbidden with a long lifetime, the lifetime
at low temperature, which corresponds to τ1 , represents the maximum contribution
of the 2 E state observed in the 25-300 K temperature range; however it cannot be
assumed that τ1 is the intrinsic 2 E lifetime. The observed thermal quenching of the
lifetime is due to thermal activation of the 4 T2 energy level contribution.
Figure 5.60: Pressure dependence of the GGG: 0.5%Cr3+ nanocrystals RT luminescence upon excitation
at 21980 cm−1 .
• High pressure effect.
RT emission spectra in GGG: 0.5%Cr3+ upon excitation at 21980 cm−1 for vari-
ous pressures up to 20 GPa are shown in Fig. 5.60. At 1.5 GPa, the broadband
luminescence from the 4 T2 Cr3+ state with a small 2 E emission contribution is ob-
served. When pressure is raised the emission changes, and above 15 GPa the band
Figure 5.61: Schematic representation of the configurational coordinate curves in harmonic approxi-
mation, Q(a1g ), for the 4 T2 and 2 E Cr3+ emitting states at different pressures.
due to 2 E → 4 A2 Cr3+ emission is predominant. Nevertheless, a structured narrow

R-line emission is not obtained for GGG: 0.5%Cr3+ nanoparticles at RT even at
pressures as high as 20 GPa. In this sense, the pressure-induced changes in GGG:
Cr3+ nanoparticles correspond neither with the results obtained for the bulk [69],
nor with the temperature dependence of the Cr3+ luminescence described in the
previous section. A slight red-shift with pressure is observed for the ruby-like emis-
sion. The effect of pressure on the 4 T2 and 2 E states can be described considering
that the 4 T2 energy depends linearly on the crystal-field splitting, ∆, while the 2 E
state is weakly coupled to the lattice. When pressure increases, the volume of the
CrO6 octahedron decreases, and therefore the bond distances RCr-O decrease; since
∆ ∝ 1/R5 , the 4 T2 energy increases. On the contrary, the 2 E energy is only weakly
affected by pressure by the small decrease of B [51]. This fact leads to an important
increment of the energy difference between both states, ∆E, with pressure (Fig.
5.61), and to a consequent decrease in the contribution of the 4 T2 state to the mixed
emitting state Ψ1 , reflected in the reduction of the d mixing coefficient. As a result,
the spin-allowed character of the emission is reduced. Besides, the thermal popu-
lation of the 4 T2 state is also reduced due to the increase of ∆E upon increasing
pressure.
The pressure dependence of the emission lifetimes in GGG: 0.5%Cr3+ nanocrystals

has been measured at two different zones centered at 14410 cm−1 and 13985 cm−1 ,
after pulsed excitation at 21980 cm−1 , and the best exponential fittings are collected
in Table 5.9. The fact that the same lifetimes have been measured at both energies
reflects again the coupling between the 4 T2 and 2 E states. The Cr3+ emission
evolution when pressures increases, from a broadband 4 T2 → 4 A2 luminescence to
a ruby-like 2 E → 4 A2 emission, is accompanied, as expected, by an increase in the
emission lifetime. Similar behavior although with longer lifetimes was observed for
the bulk in Ref. [69]. The reduction of both the thermal population of the 4 T2 state,
and the 2 E-4 T2 spin orbit mixing contribute to the emission lifetime increase with
pressure.
Table 5.9: Lifetimes of the GGG: 0.5%Cr3+ nanocrystals emission in the 1.5-20 GPa
range.
P / GPa τ /ms τ /ms

(14410 cm−1 ) (13985 cm−1 )
1.0 0.24 ± 0.02 0.22 ± 0.02
1.5 0.24 ± 0.02 0.25 ± 0.02
4.0 0.45 ± 0.05 0.40 ± 0.06
7.6 1.03 ± 0.07 0.96 ± 0.03
10.0 1.87 ± 0.06 1.79 ± 0.04
14.6 3.00 ± 0.08
20.1 3.40 ± 0.10
Since the energy separation between both states, ∆E, increases with pressure (see
Fig. 5.61), it is possible to model the dependence of lifetime on pressure assuming
a linear pressure dependence: ∆EP = ∆E + αP . Figure 5.62 shows the pressure
dependence of the Cr3+ emission lifetime in GGG: Cr3+ nanoparticles as well as the
fitting to eq. 5.9. The resultant fitting values are α = 74 cm−1 /GPa, τ1 = 3.51 ms
and τ2 = 29.5 µs. The fitted shift rate is consistent with the experimental values
observed in bulk GGG: Cr3+ (α = 106 cm−1 /GPa) [69], and in Cr3+ -doped LiCaAlF6
(α = 86 cm−1 /GPa) [67]. Since a steady-state lifetime seems to be reached in Fig.
5.62 at 20 GPa, τ1 and τ2 are an accurate estimation for the lifetimes of the 2 E and
4
T2 states, respectively. The fact that the τ2 value is significantly smaller than the
pure 4 T2 lifetime previously reported for bulk GGG: Cr3+ (65 µs) and YAG: Cr3+
(136 µs) is remarkable [69].
Figure 5.62: Pressure dependence of the Cr3+ luminescence lifetime in GGG: Cr3+ nanocrystals. The
line corresponds to the least-square fitting to eq. 5.9 (see text).
Conclusions
A study on the temperature and pressure dependence of the luminescence and lifetime
properties of Cr3+ -doped GGG nanoparticles (30 nm in size) has been carried out. A
spectral transformation from broadband, 4 T2 → 4 A2 , to narrowband, 2 E → 4 A2 emission,
along with an increase in the lifetime is measured for GGG nanoparticles at low tempera-
tures. This behavior with temperature is analogous to the pressure dependence observed
in GGG: Cr3+ bulk and it indicates that low temperature may have the same tuning
effects on nanoparticles as high pressure in the bulk. On the contrary, the influence of
high pressure on GGG: Cr3+ nanoparticles luminescence is slightly different. A change
from Cr3+ broadband to ruby-like emission is also detected, but at RT structured R-line
emission is not obtained even at 20 GPa. The lifetime dependence on temperature and
pressure has been modeled considering the spin-orbit coupling between the 4 T2 and 2 E
Cr3+ states.
Bibliography
[1] S. Heer, K. Kömpe, H.U. Güdel, and M. Haase. Highly Efficient Multicolor Upcon-
version Emission in Transparent Colloids of Lanthanide-Doped NaYF4 Nanocrystals.
Adv. Mater., 16: 2102–2105, 2004.
[2] K.W. Krämer, D. Biner, G. Frei, H.U. Güdel, M.P. Hehlen, and S.R. Lüthi. Hexag-
onal Sodium Yttrium Fluoride Based Green and Blue Emitting Upconversion Phos-
phors. Chem. Mater., 16: 1244–1251, 2004.
[3] M.F. Joubert. Photon avalanche upconversion in rare earth laser materials. Opt.
Mater., 11: 181–203, 1999.
[4] E. Heumann, S. Bär, K. Rademaker, G. Huber, S. Butterworth, A. Diening, and

W. Seelert. Semiconductor-laser-pumped high-power upconversion laser. Appl. Phys.
Lett., 88: 061108, 2006.
[5] S. Sivakumar, P.R. Diamente, and F.C.J.M. van Veggel. Silica-Coated Ln3+ -Doped
LaF3 Nanoparticles as Robust Down- and Upconverting Biolabels. Chem.-A Eur. J.,
12: 5878–5884, 2006.
Lett., 8: 3834–3838, 2008.
[7] T. Trupke, A. Shalav, B.S. Richards, P. Würfel, and M.A. Green. Efficiency enhance-
ment of solar cells by luminescent up-conversion of sunlight. Solar Energy Mater.
Solar Cells, 90: 3327–3338, 2006.
141
142 BIBLIOGRAPHY
[9] J.C. Boyer, F. Vetrone, J.A. Capobianco, A. Speghini, and M. Bettinelli. Yb3+ ion as
a sensitizer for the upconversion luminescence in nanocrystalline Gd3 Ga5 O12 :Ho3+ .
Chem. Phys. Lett., 390: 403–407, 2004.
[10] F.W. Ostermayer and L.G. Van Uitert. Cooperative Energy Transfer from Yb3+ to
Tb3+ in YF3 . Phys. Rev. B, 1: 4208–4212, 1970.
[11] R.S. Brown, W.S. Brocklesby, W.L. Barnes, and J.E. Townsend. Cooperative energy
transfer in silica fibres doped with ytterbium and terbium. J. Lumin., 63: 1–7, 1995.
[12] G.M. Salley, R. Valiente, and H.U. Güdel. Cooperative Yb3+ -Tb3+ dimer excitations
and upconversion in Cs3 Tb2 Br9 : Yb3+ . Phys. Rev. B, 67: 134111, 2003.
[13] W. Strek, P.J. Dereń, A. Bednarkiewicz, Y. Kalisky, and P. Boulanger. Efficient up-
conversion in KYb0.8 Eu0.2 (WO4 )2 crystal. J. Alloys Compounds, 300-301: 180–183,
2000.
[14] G.S. Maciel, A. Biswas, and P.N. Prasad. Infrared-to-visible Eu3+ energy upconver-
sion due to cooperative energy transfer from an Yb3+ ion pair in a sol-gel processed
multi-component silica glass. Opt. Commun., 178: 65–69, 2000.
[15] R.W.G. Wyckoff, editor. Crystal Structures, volume 2. John Wiley and Sons, New
York, London, 1963.
[16] M. Mandel. Paramagnetic Resonance of Yb3+ in Yttrium Oxide. Appl. Phys. Lett.,
2: 197–198, 1963.
Er3+ . J. Phys. Chem. B, 106: 1181–1187, 2002.
[18] R. Buisson and J.C. Vial. Transfer inside pairs of Pr3+ in LaF3 studied by up-
conversion fluorescence. J. Phys. (France) Lett., 42: 115–118, 1981.
BIBLIOGRAPHY 143
[19] R. Valiente, O.S. Wenger, and H.U. Güdel. Upconversion luminescence in Yb3+ doped
CsMnCl3 : Spectroscopy, dynamics, and mechanisms. J. Chem. Phys., 116: 5196–
5204, 2002.
[20] J.F. Suyver, J. Grimm, M.K. van Veen, D. Biner, K.W. Krämer, and H.U. Güdel.
Upconversion spectroscopy and properties of NaYF4 doped with Er3+ , Tm3+ and/or
Yb3+ . J. Lumin., 117: 1–12, 2006.
[21] A. Aebischer, M. Hostettler, J. Hauser, K. Krämer, T. Weber, H.U. Güdel, and H.B.
Bürgi. Structural and Spectroscopic Characterization of Active Sites in a Family of
Light-Emitting Sodium Lanthanide Tetrafluorides. Angew. Chem. Int. Ed., 45: 2802–
2806, 2006.
[22] P. Boutinaud, R. Mahiou, N. Martin, and M. Malinowski. Luminescence from Pr3+

3
P1 and 3 P2 states in β-NaYF4 : Pr3+ . J. Lumin., 72-74: 809–811, 1997.

Phys. Rev. B, 61: 3337–3346, 2000.
2005.
[26] J.C. Boyer, F. Vetrone, J.A. Capobianco, A. Speghini, M. Zambelli, and M. Bettinelli.
Investigation of the upconversion processes in nanocrystalline Gd3 Ga5 O12 : Ho3+ . J.
Lumin., 106: 263–268, 2004.
[27] M. Daldosso, D. Falcomer, A. Speghini, M. Bettinelli, S. Enzo, B.Lasio, and S. Polizzi.

Synthesis, structural investigation and luminescence spectroscopy of nanocrystalline
Gd3 Ga5 O12 doped with lanthanide ions. J. Alloys Compounds, 451: 553–556, 2008.
144 BIBLIOGRAPHY
[28] M. Pang and J. Lin. Growth and optical properties of nanocrystalline Gd3 Ga5 O12 :
Ln (Ln= Eu3+ , Tb3+ , Er3+ ) powders and thin films via Pechini sol-gel process. J.
Cryst. Growth, 284: 262–269, 2005.

[30] B. Keszei, J. Paitz, J. Vandlik, and A. Süveges. Control of Nd and Cr concentra-

tions in Nd, Cr: Gd3 Ga5 O12 single crystals grown by Czochralski method. J. Cryst.
Growth, 226: 95–100, 2001.
[31] P. Dorenbos. The 4fn ↔ 4fn−1 5d transitions of the trivalent lanthanides in halogenides
and chalcogenides. J. Lumin., 91: 91–106, 2000.
[32] G. Blasse and A. Bril. Philips Res. Repts, 22: 481, 1967.
[33] A. Meijerink, R. Wegh, P. Vergeer, and T. Vlugt. Photon management with lan-
thanides. Opt. Mater., 28: 575581, 2006.
[34] B. Lai, J. Wang, and Q. Su. Ultraviolet and visible upconversion emission in
Tb3+ /Yb3+ co-doped fluorophosphate glasses. Appl. Phys. B, 98: 4147, 2010.
[35] H. Liang, G. Chen, L. Li, Y. Liu, F. Qin, and Z. Zhang. Upconversion luminescence in
Tb3+ /Yb3+ -codoped monodisperse NaYF4 nanocrystals. Opt. Commun., 282: 3028–
3031, 2009.
[36] S. Sivakumar and F.C.J.M. van Veggel. Red, Green, and Blue Light Through Co-
operative Up-Conversion in Sol-Gel Thin Films Made with Yb0.80 La0.15 Tb0.05 F3 and
Yb0.80 La0.15 Eu0.05 F3 Nanoparticles. J. Disp. Technol., 3: 176–183, 2007.
[37] W.F. van der Weg, Th.J.A. Popma, and A.T. Vink. Concentration dependence of UV
and electron-excited Tb3+ luminescence in Y3 Al5 O12 . J. Appl. Phys., 57: 5450–5456,
1985.
BIBLIOGRAPHY 145
[38] H. Wang, C. Duan, and P.A. Tanner. Visible Upconversion Luminescence from Y2 O3 :
Eu3+ , Yb3+ . J. Phys. Chem. C, 112: 16651–16654, 2008.
[39] F.S. Liu, Q.L. Liu, J.K. Liang, J. Luo, L.T. Yang, G.B. Song, Y. Zhang, L.X. Wang,
J.N. Yao, and G.H. Rao. Crystal structure and photoluminescence of Tb3+ doped
Y3 GaO6 . J. Alloys Compd., 425: 278–283, 2006.
[40] Y.P. Naik, M. Mohapatra, N.D. Dahale, T.K. Seshagiri, V. Natarajan, and S.V.
Godbole. Synthesis and luminescence investigation of RE3+ (Eu3+ , Tb3+ and Ce3+ )-
doped lithium silicate (Li2 SiO3 ). J. Lumin., 129: 1225–1229, 2009.
[41] X. Liu, X. Wang, and Z. Wang. Selectively excited emission and Tb3+ -Ce3+ energy
transfer in yttrium aluminum garnet. Phys. Rev. B, 39: 10633–10639, 1989.
[42] R.S. Meltzer, S.P. Feofilov, B. Tissue, and H.B. Yuan. Dependence of fluorescence
lifetimes of Y2 O3 : Eu3+ nanoparticles on the surrounding medium. Phys. Rev. B,
60: R14012–R14015, 1999.
[43] A. Speghini, M. Bettinelli, P. Riello, S. Bucella, and A. Benedetti. Preparation,

structural characterization, and luminescence properties of Eu3+ -doped nanocrys-
talline ZrO2 . J. Mater. Res., 20: 2780–2791, 2005.
[44] G.M. Salley, R. Valiente, and H.U. Güdel. Phonon-assisted cooperative sensitization
of Tb3+ in SrCl2 : Yb, Tb. J. Phys.: Condens. Matter, 14: 5461–5475, 2002.
[45] K. Krämer and H.U. Güdel. Upconversion luminescence in K2 LaX5 : Er3+ (X = Cl,
Br). J. Alloy. Compd., 207-208: 128–132, 1994.
[46] O.S. Wenger, G.M. Salley, R. Valiente, and H.U. Güdel. Luminescence upconversion
under hydrostatic pressure in the 3d-metal systems Ti2+ : NaCl and Ni2+ : CsCdCl3 .
Phys. Rev. B, 65: 212108, 2002.
146 BIBLIOGRAPHY
2002.
[49] R. Valiente, O.S. Wenger, and H.U. Güdel. Near-infrared-to-visible photon upcon-
version process induced by exchange interactions in Yb3+ -doped RbMnCl3 . Phys.
Rev. B, 63: 165102, 2001.
[50] P. Gerner, O.S. Wenger, R. Valiente, and H.U. Güdel. Green and Red Light Emission
by Upconversion from the near-IR in Yb3+ Doped CsMnBr3 . Inorg. Chem., 40: 4534,
2001.
[52] S.C. Abrahams, P. Marsh, and C.D. Brandle. Laser and phosphor host
La1−x MgAl11+x O19 (x = 0.050): Crystal structure at 295 K. J. Chem. Phys.,
86: 4221–4227, 1987.
[53] S. Nukuta and T. Onimaru. Patent 7.037.445. 2006.
[54] V. Adelskold. X-ray studies on magneto-plumbite and other substances resembling

beta-alumina. Arkiv. Kemi Mineral. Geol. A 12, 29: 1–9, 1938.
[55] N. Iyi, Z. Inoue, S. Takekawa, and S. Kimura. The crystal structure of lanthanum
hexaaluminate. J. Solid State Chem., 54: 70–77, 1984.
[56] M. Gasperin, M.C. Saine, A. Kahn, F. Laville, and M. Lejus. Influence of M2+ Ions
Substitution on the Structure of Lanthanum Hexaaluminates with Magnetoplumbite
Structure. J. Solid State Chem., 54: 61–69, 1984.

Lett., 25: 1710–1712, 1970.
J. Chem. Phys., 98: 1768–1775, 1993.
BIBLIOGRAPHY 147
[59] R.T. Wegh and A. Meijerink. Cooperative luminescence of ytterbium(III) in La2 O3 .

Chem. Phys. Lett., 246: 495–498, 1995.

[61] M.N. Sanz-Ortiz and F. Rodrı́guez. Photoluminescence properties of Jahn-Teller

transition-metal ions. J. Chem. Phys., 131: 124512, 2009.
[63] J. Luzón, J. Campo, F. Palacio, G.J. McIntyre, and A. Millán. Understanding mag-
netic interactions in the series A2 FeX5 · H2 O (A=K, Rb; X=Cl, Br). I. Spin densities
by polarized neutron diffraction and DFT calculations. Phys. Rev. B, 78: 054414,
2008.
[64] M.P. Hehlen, A. Kuditcher, S.C. Rand, and S.R. Lüthi. Site-Selective, Intrinsically
Bistable Luminescence of Yb3+ Ion Pairs in CsCdBr3 . Phys. Rev. Lett., 82: 3050–
3053, 1999.
[66] W.A. Wall, J.T. Karpick, and B. Di Bartolo. Temperature dependence of the vibronic
spectrum and fluorescence lifetime of YAG: Cr3+ . J. Phys. C: Solid St. Phys., 4: 3258–
3264, 1971.
[67] M.N. Sanz-Ortiz, F. Rodrı́guez, I. Hernández, R. Valiente, and S. Kück. Origin of

the 2 E - 4 T2 Fano resonance in Cr3+ -doped LiCaAlF6 : Pressure-induced excited-state
crossover. Phys. Rev. B, 81: 045114, 2010.
[68] Y.R. Shen and K.L. Bray. Effect of pressure and temperature on the lifetime of Cr3+
in yttrium aluminum garnet. Phys. Rev. B, 56: 10882–10891, 1997.
148 BIBLIOGRAPHY
[69] U. Hömmerich and K.L. Bray. High-pressure laser spectroscopy of Cr3+ :

Gd3 Sc2 Ga3 O12 and Cr3+ : Gd3 Ga5 O12 . Phys. Rev. B, 51: 12133–12141, 1995.
[70] F.M. Hashmi, R.C. Powell, and G. Boulon. Energy transfer and radiationless relax-
ation processes in Nd3+ and Cr3+ doped mixed garnet crystals. Opt. Mat., 1: 281–298,
1992.
[71] J.F. Dolan, L.A. Kappers, and R.H. Bartram. Pressure and temperature dependence
of chromium photoluminescence in K2 NaGaF6 : Cr3+ . Phys. Rev. B, 33: 7339–7341,
1986.
[72] S. Heer, M. Wermuth, K. Krämer, D. Ehrentraut, and H.U. Güdel. Up-conversion

excitation of sharp Cr3+ 2 E emission in YGG and YAG codoped with Cr3+ and Yb3+ .
J. Lumin., 94-95: 337–341, 2001.
[73] E.V Zharikov, S.V Lavrishchev, V.V Laptev, V.G Ostroumov, Z.S. Saidov, V.A.
Smirnov, and I.A Shcherbakov. New possibilities for Cr3+ ions as activators of the
active media of solid-state lasers. Sov. J. Quantum Electron., 14: 332–336, 1984.
[74] B. Di Bartolo, editor. 27th Workshop: Luminescence of Inorganic Materials and

Bioimaging: Metal-to-Metal Energy and Electron Transfer. Erice, Sicily, Italy, 22-28
June 2010.
[76] F.S. Ham. Dynamical Jahn-Teller Effect in Paramagnetic Resonance Spectra: Or-
bital Reduction Factors and Partial Quenching of Spin-Orbit Interaction. Phys. Rev.,
138: A1727–A1740, 1965.
Chapter 6
Semiconductor nanoparticles.
Optical spectroscopy at high
pressure
6.1 Introduction
In this chapter, we show the main results concerning optical spectroscopy of CdS and
Zn1−x Cox O nanoparticles at high pressure and the influence of pressure-induced phase
transitions on these properties. The physical properties in the nanocrystal regime can
be quite different from those of the bulk. For instance, there are many works which
demonstrate a decrease in the melting temperature [1], [2], [3] as well as an increase of
the transition pressure [4], [5] with decreasing nanoparticles size. This behavior may
arise because of the relatively high energy surfaces in nanocrystals which make the high
pressure phase less stable than in the bulk [4]. Besides, nanoparticles lack the internal
defects that serve as nucleation points for structural phase transitions [6].
CdS is a well-studied II-VI semiconductor due to its many applications such as bio-
labeling or light emitting diodes [7], [8]. The bulk crystallizes in the W structure while
both W and RS structures have been found for CdS QDs [9]. The main interest in the
study of its optical properties arises from the fact that they are size-dependent. Ab-
sorption, Raman and XRD measurements have been performed in this work in order to
characterize the ZB-to-RS phase transition in CdS nanoparticles prepared by ball milling.
DMSs are currently attracting intense interest in different applied areas like the emerg-
ing field of spin-based electronics [10], [11]. Although TM2+ -doped ZnO DMSs have been
149
150 6.2. CdS nanoparticles
proposed to be favorable candidates for RT ferromagnetism, the magnetic behavior is

extremely sensitive to the synthesis procedure and remains controversial [12], [13]. The
magnetization measurements carried out by M. A. White et al. on W-Zn1−x Cox O colloidal
nanocrystals prepared following the same synthesis method as in this work, indicated a
paramagnetic-like magnetism at low Co2+ concentrations and an antiferromagnetic be-
havior for W-CoO, with no ferromagnetic response [14]. In our case, Co2+ impurities
have been used as a local probe to monitor the phase transition under high pressure in
Zn1−x Cox O nanocrystals. The RS-to-W phase transition has also been studied by means
of bandgap absorption and Raman spectroscopy.
6.2 CdS nanoparticles

6.2.1 Synthesis and characterization
Depending on the particle size, different structures have been identified for CdS nanocrys-
tals [9]. (i) When the size is smaller than 5 nm, CdS nanoparticles are mainly crystallized
in the metastable cubic ZB structure (space group F 4̄3m). (ii) When the size is larger
than about 8 nm, they are in the stable hexagonal W structure (P 63 mc space group).
(iii) In between, they are in intermediate phases. In both ZB and W structure each an-
ion is surrounded by four cations at the corners of a tetrahedron and viceversa. These
results have also been observed in the structural analysis of colloidal CdSe nanocrystals
deduced from XRD by Bawendi et al. [15]. In bulk CdS crystals with the W structure,
the ΓV15 → ΓC
1 direct gap is the lowest in energy (Eg = 2.4 eV at 300 K) and indirect
gaps occur at least 1 eV above Eg and always stay above Eg at high pressure [16]. Inves-
tigations of electronic structure in W- and ZB-CdS showed that the fundamental direct
bandgaps in both structures differ by less than 0.1 eV [17]. Since the exciton Bohr radius
in bulk CdS is 2.6 nm [18], CdS nanocrystals with the metastable ZB structure (diameter
less or equal to 5 nm) are in the weak confinement regime. Therefore, the direct optical
energy gaps of bulk and nanocrystals are similar.
Pure and Yb3+ -doped CdS nanocrystals have been prepared in a planetary ball mill as
described in Section 3.2.3. These materials show an intense orange color. The initial idea
6. Semiconductor nanoparticles. Optical spectroscopy at high pressure 151
Figure 6.1: XRD pattern of the CdS nanocrystalline samples prepared by ball milling. The black line
corresponds to the calculated pattern for the ZB-W mixed structure, and ticks indicate the (hkl) index
for each crystallographic phase.
for this system was obtaining bandgap UC emission upon Yb3+ excitation. XRD pattern
of CdS nanoparticles together with a Rietveld refinement considering both cubic and
hexagonal phases is shown in Fig. 6.1. The best fitting for the prepared nanocrystalline
sample is obtained for a 85% ZB - 15% W mixed structure; however, the error associated
to this quantitative analysis is pretty large, and variations in some fitting parameters may
lead the ZB contribution to a minimum value of 65%. The cell parameter obtained for
the ZB phase, a=5.7844 Å, is slightly smaller than the bulk value, a0 =5.818 Å [19]. The
values for the W structure, a=4.1111 Å and c=6.5837 Å, are somewhat different to those
c0
of the bulk (a0 =4.137 Å and c0 =6.7144 Å) [20]. The ac (nano) = 1.60 and a0
(bulk) = 1.62
ratios give a 1% deviation from the bulk; this estimation of the W structure distortion is
smaller than the up to 15% differences observed by Kumpf et al. [21].
The average particle size of the as-obtained nanoparticles, according to the Williamson-
Hall equation (eq. 4.1), has been determined to be about 5 nm. This value is in good
agreement with the TEM results shown in Fig. 6.2. However, TEM images also show that
the samples are made of particles of different shapes and sizes, from spherical nanoparticles
of about 5 nm in size to bigger particles and nanorods of about 3 − 5 nm in diameter.
Figure 6.2: TEM images of Yb3+ -doped CdS nanocrystals prepared by ball milling.
6.2.2 Optical properties and X-ray diffraction under high pres-

sure
The experimental results of the absorption, Raman and XRD experiments on CdS nanopar-
ticles under high pressure have been analyzed considering only the ZB structure at low
pressures, and are presented in this section.
Absorption
Some representative CdS nanocrystals absorption spectra at different pressures are shown
in Fig. 6.3. An analysis of the spectra at low pressures indicates that the absorption at
the fundamental edge can be described by eq. 2.19. This corresponds to a direct gap
transition between parabolic bands, VB and CB, and it is in agreement with the band
structure of the metastable ZB structure of CdS at AP [16], [17]. When the pressure is
raised above 6 GPa, the CdS nanocrystals suddenly become brown colored, and a large
and abrupt red-shift is observed in the absorption edge. Previous high pressure optical
studies in bulk CdS indicated an abrupt red-shift in the optical absorption edge at 2.7-3.0
GPa, and this was identified to be due to a first-order W-to-RS phase transition [22],
[23]. Moreover, the transition pressure for the ZB-to-RS phase transition was found to
be around 8 GPa in CdS colloidal nanocrystals [24]. In this work, the ZB-to-RS phase
transition is also observed at pressures much higher than the bulk transition pressure of ca.
3 GPa. This kind of result has been usually observed in nanocrystalline semiconductors
systems of the group IV as well as III-V or II-VI in comparison with the bulk counterpart
[25]. The high transition pressure can be explained by a higher value of the surface tension
for the RS phase nanocrystals compared to the ZB-CdS [6]. An example of the absorption
Figure 6.3: Shift in the absorption spectra of CdS nanoparticles with hydrostatic pressure.
edge for the high pressure RS phase recorded at 6.1 GPa is shown in Fig. 6.3. The edge
in the RS phase is strikingly different compared to that of CdS nanocrystals in the ZB
phase, and is clearly due to an indirect transition.
The ZB-CdS optical energy bandgap values, Egd (ZB), have been obtained by extrapo-
lating the square of the absorbance fitting curves, α2 versus ~ω, to α = 0. The resulting
Egd (ZB) pressure dependence is shown in Fig. 6.4, and a blue-shift of the direct bandgap
Figure 6.4: Pressure dependence of the optical energy bandgap, Eg , in both ZB and RS structures for
CdS nanocrystals.
with pressure is observed. For most semiconductors, the variation of the direct bandgap
under pressure can be described by the following quadratic expression [24]:
Eg (P ) = Eg (0) + aP + bP 2 (6.1)
The fit of the experimental data to eq. 6.1 gives the following values: Eg (0) = 2.33
eV, a = 31 ± 4 meV/GPa and b = −1.1 ± 0.3 meV/GPa2 . The obtained linear pressure
coefficient is lower than the one reported for the direct energy gap for ZB-CdS colloidal
nanoparticles, 45.7 meV/GPa [24], and for bulk W-CdS, 45.5 ± 0.5 meV/GPa [23]. In
Fig. 6.4, the estimated indirect energy gap in the RS phase, Egi (RS), is also plotted as a
function of pressure. The Egi (RS) values have been obtained from a plot of α1/2 versus
~ω − Egi and extrapolating the resultant straight line to α = 0. The fitted linear pressure
dependence is −35 ± 5 meV/GPa. A similar negative shift was observed in II-VI analog
semiconductors in the RS phase at high pressure [26]. However, it is very difficult to
determine exactly the indirect bandgap of CdS nanoparticles in the high pressure RS
phase due to the formation of band-tail states because of the large number of defects and
dislocations induced by pressure throughout the phase transitions. The highest pressure
achieved in the up-stroke is 11 GPa. Absorption spectra has also been recorded upon
releasing pressure down to 1.6 GPa. As observed in Fig. 6.4, the bandgap does not
revert to its original value in the ZB phase within this pressure range. This behavior is
clearly opposite to the reversible transition with no hysteresis observed in CdS colloidal
nanocrystals by Haase and Alivisatos [24].
Raman spectroscopy
The CdS nanoparticles Raman spectra have been studied at RT under hydrostatic pressure
up to 7 GPa (Fig. 6.5). The LO and 2LO modes are observed at 304 and 608 cm−1 at
AP, and shift linearly with pressure to higher energies (see Fig. 6.6(a)) with a pressure
coefficient of 4.8 ± 0.2 cm−1 /GPa for the LO mode, and 9.2 ± 0.4 cm−1 /GPa for the
2LO one, respectively. These slope values are in good agreement with previous results
for bulk W-CdS [27], [28] and ZB-CdS colloidal nanocrystals [24]. In addition to these
modes, a shoulder is observed overlapping the LO mode, and it has been ascribed to a
Figure 6.5: RT Raman spectra (λexc = 514.5 nm) of CdS nanoparticles at different pressures (up-
stroke).
surface mode [29]. In nanocrystalline samples the contribution of the surface scattering
may be comparable to the volume Raman scattering; Scott and Damen also indicated
the observation of surface modes in CdS by Raman spectroscopy [30]. The dependence of
the surface mode energy on pressure is shown in Fig. 6.6(b). Above 6.8 GPa the Raman
peaks disappears proving the ZB-to-RS CdS nanocrystals phase transition.
The Raman spectra of the CdS nanocrystals have been recorded after the phase tran-
sition by decreasing pressure from 8 to 0 GPa. The data are shown in Fig. 6.7(a). It
can be seen that no Raman peaks are recovered down to 1.3 GPa and hence, ZB phase
remains within this pressure range. Nevertheless, LO and 2LO Raman modes appear
again when pressure is released to 0 GPa; therefore, it is concluded that the ZB-to-RS
phase transition is reversible although presents a large hysteresis. Figure 6.7(b) compares
the CdS nanoparticles Raman spectra at AP before and after the phase transition. The
Figure 6.6: Pressure dependence of the energy of the observed phonons, Eph , for CdS nanocrystals.
LO and 2LO Raman modes (a), and surface mode (b).
fact that both spectra are centered at the same frequency indicates that no appreciable
change in size is induced by the phase transition.
Figure 6.7: RT Raman spectra (λexc = 514.5 nm) of CdS nanoparticles upon decreasing pressure (a)
and comparison between the AP Raman spectra before and after the pressure cycle (b).
The variation of the intensity of LO and 2LO Raman modes with pressure is depicted
in Fig. 6.8. An intensity reduction is observed upon increasing pressure. According to the
Raman cross-section (eq. 2.10), the peaks intensity should be compensated for possible
absorption effects at the bandgap to quantify the real Raman intensity decrease due to
the ZB-to-RS phase transition onset. However, since in our particular case the excitation
energy, 2.41 eV (514.5 nm), is close to the bandgap values, the bandgap absorption
contribution cannot be estimated.
Figure 6.8: Pressure dependence of the normalized intensity of LO and 2LO Raman modes for CdS
nanocrystals.
X-ray diffraction
Figure 6.9 shows the evolution of the CdS nanoparticles XRD patterns under high pressure
up to 7.5 GPa in both the up and down stroke. It is observed that peaks corresponding to
ZB-CdS shift to higher angles with increasing pressure, consistent with a volume decrease.
This phase starts to transform into the RS structure at around 4.3 GPa, and both phases
coexist between 4.3 and 5.4 GPa. Above this pressure, only the RS structure is observed.
The RS phase is maintained during the whole releasing pressure process, and therefore,
from XRD experiments we would say that this phase transition is irreversible. Preceding
XRD studies showed that bulk W-CdS crystals transformed to the RS structure between
2 and 3 GPa [31], [32].
The cell parameters at different pressures have been determined from the XRD pat-
terns Rietveld refinement. Only ZB and RS phases have been considered before and after
the phase transition respectively, while both structures have been included in the fittings
during the phases coexistence, between 4.3 and 5.4 GPa. The volume dependence on
pressure has been fitted to a Murnaghan equation of state (eq. 2.23) and is depicted in
Fig. 6.10. A 7% volume reduction is observed between 0 and 5 GPa, besides, a 15%
volume change has been detected at the phase transition.
Figure 6.9: XRD patterns of CdS nanoparticles under high pressure. Asterisks represent the diffraction
peaks of the hydrostatic medium (silicon oil).
The following bulk modulus values have been obtained by fixing its pressure derivative
to B00 = 4; B0 = 74±2 GPa for the ZB structure, and B0 = 107±5 GPa for the RS phase.
Therefore, the ZB structure is more compressible than the RS one. The bulk modulus
values for the CdS single crystals are around 60 GPa in the W phase [32], [33], and ca.
85 GPa in the RS structure [32], [34]. An increase in both the relative volume change
at the transition pressure, and the bulk modulus value when decreasing particle size was
previously described in Fe2 O3 nanocrystals [35], or Ag and Au nanoparticles [36], and it
was ascribed to surface effects or bond length compression within the particles due to
strain effects.
Figure 6.10: Isothermal equation of state of CdS nanoparticles for the ZB (red circles) and RS (black
circles in the up-stroke and blue circles in the down-stroke) structures. In both cases the volume depen-
dence on pressure has been fitted to a Murnaghan equation.
6.2.3 Conclusions
CdS nanocrystals in the metastable ZB cubic structure with an average diameter of 5 nm

have been prepared in a planetary ball mill. The ZB-to-RS phase transition has been
identified by changes in the optical absorption and Raman spectra as well as in the XRD
patterns. The direct optical energy gap of ZB-CdS nanoparticles increases with pressure
and the value of the linear pressure coefficient is a = 31±4 meV/GPa. When the pressure
is raised above 6 GPa, a change from direct to indirect gap occurs. In the RS phase, the
pressure dependence of the indirect gap is −35 ± 5 meV/GPa. The energy of LO and 2LO
Raman modes for CdS nanoparticles increases with pressure with the same rate as bulk
W-CdS. Raman measurements have also evidenced that the ZB-to-RS phase transition for
CdS nanocrystals takes place at around 6.0-6.5 GPa, and the fact that the ZB structure
is recovered at AP in the down-stroke. XRD measurements under high pressure have
shown that ZB-CdS nanoparticles undergo a transition into the RS structure above 5.4
GPa. Bulk modulus of B0 = 74 ± 2 GPa and B0 = 107 ± 5 GPa have been obtained for
the low-pressure and high-pressure structures, respectively. It has been observed in three
different experiments that the stability domain of the ZB structure increases up to 6 GPa
at RT for CdS nanocrystals with 5 nm diameter.
160 6.3. Colloidal nanocrystalline Zn1−x Cox O
6.3 Colloidal nanocrystalline Zn1−xCoxO

6.3.1 Synthesis and characterization
ZnO crystallizes at RT and AP in the W-type structure (hexagonal, P 63 mc space group).

This structure is formed from a hexagonal Bravais lattice and a motive of two atoms of
each element. The lattice parameters are a0 =3.2495 Å and c0 =5.2069 Å [37]. In an ideal
W structure, cations and anions form two compacted interpenetrated hexagonal lattices.
Each atom (Zn or O) is tetrahedrally coordinated and occupies sites of C3v symmetry
(Fig. 6.11).
Figure 6.11: Zn1−x Cox O W structure. The Zn2+ (or Co2+ ) ions (red) are tetrahedrally coordinated
by oxygens (green).
The application of hydrostatic pressure to ZnO causes a decrease in the unit cell
volume and an increase in the coulomb repulsion. At a given pressure, it becomes en-
ergetically favorable for ZnO to assume the more compact RS structure (cubic, F m3̄m
space group). The RS structure may be described by two face-centre-cubic (fcc) lattices
shifted by a/2, with a0 =4.275 Å. In this case, both atoms (Zn and O) are in an octahedral
coordination symmetry (Fig. 6.12). It was demonstrated that on the pressure up-stroke,
bulk polycrystalline ZnO undergoes a W-to-RS phase transition at around 9 GPa [38],
[39]. On the down-stroke, bulk ZnO was seen to recover its original W structure at either
2 GPa or 4 GPa [39], [40].
The electronic band structure of W-ZnO at RT and AP (Fig. 6.13 left) consists of
three broad regions with the CB comprising of Zn 4s levels, the upper VB of O 2p levels
and the lower VB of Zn 3d levels. It must be pointed out that the bandgap occurs at Γ
Figure 6.12: Zn1−x Cox O RS structure. The Zn2+ (or Co2+ ) (blue) are octahedrally coordinated by
oxygens (black).
Figure 6.13: Electronic band structure of W-ZnO (left) and RS-ZnO (right) taken from Ref. [43].
and it is a direct transition (Eg = 3.4 eV for bulk ZnO) [41]. The band structure of the
RS phase (Fig. 6.13 right) indicates that the interband transition is no longer a direct gap
at Γ, but it becomes an indirect gap between L and Γ, with reference to the fcc Brillouin
zone [42].
ZnO colloidal nanocrystalline samples doped with different Co2+ concentrations have
been prepared in this work according to the synthesis methods described in Section 3.7.
XRD measurements carried out by M.A. White et al. (patterns shown in Fig. 6.14)
indicate a good crystallization in the W structure in all cases [14]. The cell parameters
obtained for W-ZnO nanoparticles, a=3.2078 Å and c=5.1962 Å, are slightly smaller than
the standard bulk values (a0 =3.2495 Å and c0 =5.2069 Å). This lattice contraction with
Figure 6.14: XRD patterns of Zn1−x Cox O colloidal nanocrystals carried out by M.A. White, Univ. of
Washington. The calculated pattern for W-ZnO is also shown.
decreasing particle size was previously observed in metallic nanoparticles [44].
TEM images of ZnO: 5%Co2+ colloidal nanocrystals (Fig. 6.15) show spherical nanopar-
ticles with an average particle diameter of ca. 4 nm. This size is around four times larger
than the size of the bulk exciton (0.9 nm), and therefore, we are again in the weak con-
finement regime [45]. A narrow size distribution for the W-Zn1−x Cox O nanoparticles
prepared by this method was previously demonstrated by D.A. Schwartz et al. [46].
Figure 6.15: TEM and HRTEM images of ZnO: 5%Co2+ nanoparticles.

6.3.2 Optical properties under high pressure
In this work, the RT electronic structure of W-Zn1−x Cox O nanoparticles has been inves-
tigated by means of optical absorption and Raman spectroscopy under pressure.
Absorption
• ZnO
Figure 6.16 shows the pressure dependence of the ZnO nanoparticles (average size
4 nm) absorption spectra up to 24 GPa. A monotonous blue-shift of the absorp-
tion edge is observed as pressure increases. Since the optical absorption edge in
W-ZnO colloidal nanocrystals is caused by direct absorption, it remains sharp with
increasing pressure, and its pressure dependence can be determined accurately ac-
cording to eq. 2.19. When the pressure is raised above 14 GPa, the W-to-RS phase
transition occurs, and the absorption edge changes dramatically compared to lower
pressures. Both direct and indirect energy gaps are observed for ZnO nanocrystals
in the RS structure. Although it is much more difficult in this case, the RS-ZnO
indirect energy gap can also be estimated from eq. 2.22.
Figure 6.16: Absorption spectra of ZnO nanocrystals (ca. 4 nm) upon increasing pressure.
In previous absorption experiments, an increase in the energy bandgap with pressure

was noted for both ZnO bulk crystals and thin films, and the W → RS phase
transition was observed at around 9.5 GPa [42], [47]. The phase transition in ZnO
thin films was detected as a decrease in the absorbance at the edge. Since RS-
ZnO is an indirect gap semiconductor with low absorption coefficient, only intense
direct absorption edges were detectable in transmission measurements in a 200 nm
thick film [47]. The transition from the W-ZnO bulk sample to the RS phase was
observed as a neat change in the shape of the absorbance [42]. An intense transition
at energies higher than 4.5 eV in both RS-ZnO bulk and films was also assigned to
an allowed direct transition [42], [47]. High pressure synchrotron XRD experiments
indicated the transition pressure to be around 10.5 GPa for W-ZnO with 50 nm in
diameter, and 15 GPa for 12 nm nanocrystalline ZnO [48], [49]. Grzanka et al. also
studied the size effect on the high pressure phase transition of ZnO nanocrystals,
obtaining similar results [50]. Shan et al. studied the pressure dependence of the
luminescence properties in ZnO nanowires and they found the phase transition at
around 12 GPa [51]. But, as far as we know, no experimental investigations on the
optical absorption of ZnO nanocrystals under high pressure have been reported so
far.
Figure 6.17(a) compares the variation with pressure of the direct optical energy gap
in W-ZnO colloidal nanocrystals, Egd (W), with the relative pressure dependence of
the total energy gap for the high pressure RS structure, ∆Egd+i (RS). Egd (W) values
have been obtained from the fitting of the square of the absorbance to eq. 2.19, α2 -
~ω, Egd (W) being the point at which the absorption fitting curve goes to zero. The
variation with pressure of both direct and indirect contributions to the absorption
edge in the RS phase, ∆Egd+i (RS), has been estimated considering the value of
the energy at a fixed absorbance for each pressure. In Fig. 6.17(b) the pressure
dependence of the indirect optical energy gap for the ZnO nanoparticles in the RS
structure, Egi (RS), is shown. Egi (RS) values are obtained as follows: α1/2 is plotted
versus ~ω, and the ~ω value at α = 0 is equal to Egi + ~Ω; where ~Ω is the phonon
energy, estimated to be ca. 55 meV (438.4 cm−1 ) from Raman measurements.
Figure 6.17: Pressure dependence of the ZnO direct energy gap in the W phase, Egd , and both direct
and indirect gaps in the RS structure, ∆Egd+i , (a). And pressure dependence of the direct gap for W-ZnO,
Egd , and indirect gap for RS-ZnO, Egi (b).
The pressure coefficient for the W-ZnO nanoparticles direct absorption edge is 20±2
meV/GPa (Fig. 6.17). This value is slightly lower than the one measured in W-ZnO
bulk (25 ± 2 meV/GPa) and thin film (23.0 ± 0.5 meV/GPa) [47]. For ZnO colloidal
nanocrystals in the RS structure, the relative variation of the total energy gap with
pressure is given by 70 ± 1 meV/GPa (Fig. 6.17(a)). It can be seen that the direct
energy gap increases with pressure while the indirect one decreases with a pressure
coefficient given by −130 ± 10 meV/GPa (Fig. 6.17(b)).
Absorption spectra of ZnO nanoparticles recorded in the down-stroke between 20

and 1 GPa are shown in Fig. 6.18. They indicate the non-reversibility of the W-
to-RS phase transition, that is, the metastability of RS-ZnO nanocrystals after the
pressure cycle. However, since AP has not been reached, a complete certainty of the
RS phase stability does not exist for the pure ZnO nanocrystalline sample. High
pressure structural investigations of bulk ZnO indicated that the W-to-RS transition
is reversible at RT; with a transition pressure of 9 GPa upon increasing pressure and
2 GPa in the down-stroke [39]. However, this property was already demonstrated to
be drastically different for nanocrystals, and Decremps et al. showed that metastable
RS ZnO nanoparticles were obtained upon decreasing pressure after a high-pressure
(15 GPa) and high-temperature (550 K) treatment [52].
Figure 6.18: Absorption edge of RS-ZnO nanoparticles at different pressures (down-stroke).
• ZnO: 5%Co2+
Bandgap absorption spectra of ZnO: 5%Co2+ nanoparticles for different pressures

up to 20 GPa are shown in Fig. 6.19. The dependence of the Co2+ absorption
bands upon increasing pressure is presented in Fig. 6.20. Figure 6.21 compares
the bandgap absorption in pure and 5%Co2+ -doped ZnO nanocrystals. Absorption
spectra of Zn1−x Cox O nanocrystals present three main differences with respect to
pure W-ZnO which are summarized in Fig. 6.22 [53]:
– The energy of the fundamental band-to-band absorption edge is slightly red-

shifted.
– A broad band absorption, related to a charge-transfer transition involving pro-

motion of an electron from the Co2+ to the CB, appears at energies just below
(and overlapping) the band-to-band edge [54].
– A well defined absorption band related to the spin-allowed 4 A2 → 4 T1 d−d

transition of tetrahedral Co2+ is observed in the visible region [55].
Figure 6.19: Absorption edge of ZnO: 5%Co2+ nanoparticles at different pressures (up-stroke).
Figure 6.20: RT pressure dependence of the 4 A2 → 4 T1 Co2+ transition of ZnO: 5%Co2+ nanocrystals.
Figure 6.21: Absorption edge of W-ZnO and W-ZnO: 5%Co2+ nanoparticles at AP.
An increase of the sample transparency inside the DAC is detected upon increasing
pressure, and the W-ZnO: 5%Co2+ transition to the RS phase is gradually observed
up to 15 GPa in the three absorption features:
– The fundamental absorption edge shifts to higher photon energies, and a change
of the absorption edge, when moving from direct to indirect gap scenario, is
observed.
– The charge-transfer band virtually disappears, or overlaps the direct transition

of the RS phase.
– The d−d Co2+ absorption band around 2 eV decreases its intensity by a factor
of 10 and shifts to higher energies, 2.5 eV, as a consequence of the modifica-
tion from a tetrahedral (W phase) to an octahedral (RS phase) coordination
symmetry.
Figure 6.23(a) shows the pressure dependence of the ZnO: 5%Co2+ nanocrystals
optical energy gap for the W and RS structures. Since the charge-transfer band
overlaps the bandgap absorption, the direct energy gap in the W phase, Egd (W), has
been estimated from the fitting of the square of the absorbance to eq. 2.19 like for
pure ZnO, but considering only values of the absorbance above 1.5. ∆Egd+i (RS) has
Figure 6.22: Schematic representation of the optical transitions observed in ZnO: 5%Co2+ .
been obtained from the value of the energy at a fixed absorbance for each pressure
in the RS phase. In Fig. 6.23(b) the pressure dependence of the Co2+ normalized
absorption intensity is presented. The direct bandgap for the W-ZnO: 5%Co2+
nanocrystals exhibits a linear pressure dependence with a pressure coefficient of
17 ± 2 meV/GPa (Fig. 6.23(a)). This value is lower than the one obtained in
W-ZnO: 5%Co2+ thin films (24.9 meV/GPa) [53].
Figure 6.23: Pressure dependence of the ZnO: 5%Co2+ direct energy gap in the W phase, Egd , and both
direct and indirect gaps in the RS structure, ∆Egd+i , (a). And pressure dependence of the normalized
Co2+ absorbance (b).
From Figs. 6.20 and 6.23(b) it is clearly seen that the Co2+ absorption intensity
decreases continuously, which indicates that the W-to-RS phase transition takes
place gradually, starting at pressures as low as 4-5 GPa. This result is opposite to
the large and abrupt decrease in the Co2+ absorption intensity observed in ZnO:
Co2+ single crystal by Stephens et al. [56]. The dramatic change of the absorption
edge detected in Zn1−x Cox O thin films was used to propose also a sudden W-to-
RS phase transition [53]. Nevertheless, it can be deduced from our experimental
results (see Fig. 6.23) that the pressure dependence of the bandgap absorption do
not provide information about the abrupt or continuous character of the transition.
The charge-transfer band absorbance diminution is correlated with the d−d Co2+
bands behavior.
The continuous character of the phase transition in ZnO: Co2+ nanocrystals evi-
dences that different nanoparticles within our sample present different transition
pressures, and this could be attributed to several reasons. Firstly, it could be due to
a broad size distribution, since it is well known that the transition pressure increases
for smaller particle size [4], [5]. However, this is not the case, as there are nanoparti-
cles which show the phase transition at 4 GPa, while the transition pressure is 9 GPa
for the bulk. Secondly, different orientation or morphology of the particles would
give rise to differences in the surface energy among the various nanoparticles, and
the stability under high pressure is very related to it [25]. Thirdly, if Co2+ ions are
heterogeneously distributed within our nanocrystalline sample, those nanoparticles
with higher cobalt concentration would have higher absorbance and lower transition
pressures, involving a decrease in the total absorbance. The W-to-RS phase transi-
tion was observed in CoO nanoparticles (around 50 nm in size) in the 0.8-6.9 GPa
pressure range by means of high pressure synchrotron-radiation XRD measurements
[57].
Several experiments can be accomplished in order to clarify the actual explanation

for the gradual phase transition observed in ZnO: Co2+ nanoparticles. A concentra-
tion study may be carried out on different particles at the TEM facility. Moreover,
electron diffraction experiments would give information about the structure of dif-
ferent particles. Hence, the following relevant experiment will be developed: Both
concentration and electron diffraction measurements will be performed on ZnO:
Co2+ nanocrystals after reaching an intermediate pressure in the up-stroke in order
to check if the nanoparticles with high Co2+ concentration have already undergone
Figure 6.24: Absorption edge of RS-ZnO: 5%Co2+ nanoparticles in the down-stroke from 22 GPa to
AP.
the phase transition, while those with low Co2+ concentration are still in the W
phase.
Absorption spectra of ZnO: 5%Co2+ nanoparticles recorded after the phase transi-
tion, in the down-stroke down to AP are shown in Fig. 6.24. Both the absorption
edge at AP (Fig. 6.24) which is characteristic of an indirect semiconductor, and the
fact that Co2+ tetrahedral absorption is not recovered at AP, are clear evidences of
the metastability of the RS-ZnO: 5%Co2+ phase at RT and AP.
Raman spectroscopy
The optical phonons predicted for the W-ZnO at the Γ-point of the Brillouin zone are:
Γopt = A1 + 2B1 + E1 + 2E2 . Both A1 and E1 polar modes are Raman and IR active
and split into longitudinal and transverse optical (LO and TO) components. The two
non-polar E2Low and E2High modes are Raman active and IR inactive, while the B1 silent
modes are Raman and IR inactive [58], [59]. The mode assignment for W-ZnO single
crystals at RT and AP was well established by Calleja and Cardona [60]. The phonon
confinement of ZnO nanocrystals affects both the phonon energy and the symmetry. A
red-shift in both E2Low and E2High modes as well as an asymmetric broadening on the low
energy side were observed in the Raman spectrum of ZnO nanoparticles (with 17 nm in
size) [61].
Figure 6.25: Raman spectra of ZnO: 5%Co2+ nanoparticles at different representative hydrostatic
pressures.
In previous works, the experimental phonon energies at ambient conditions for the
E2Low and E2High modes were determined to be ca. 100 and 440 cm−1 , respectively, for
both W-ZnO and W-ZnO: 5%Co2+ single crystals. The A1 (TO), E1 (TO) and E1 (LO)
modes were also clearly seen in the Raman spectra at around 380, 414 and 580 cm−1 ,
respectively. On the contrary, the A1 (LO) mode was not observed. Moreover, additional
lines in the ZnO: 5%Co2+ spectra were attributed to CoO magnetic excitations [58], [59],
[60]. Raman scattering measurements performed by Decremps et al. in bulk ZnO also
evidenced a W-to-RS structural transformation at around 9 GPa [59]. High pressure

Raman spectra of ZnO nanowires showed a reversible W-to-RS phase transition with
transition pressures of 10.3 GPa in the up-stroke and 4.4 GPa in the down-stroke [62].
The pressure dependence of the zone-center phonons (E2 , A1 and E1 ) has been mea-
sured in this work for the W-ZnO: 5%Co2+ nanoparticles and is shown in Fig. 6.25.
The E2High mode is prominently observed in the spectra while the other modes are hardly
observable and then could only be tentatively assigned. The hexagonal to cubic phase
transition is completed around 15 GPa for the W-ZnO: 5%Co2+ nanocrystals (ca. 4 nm
in size). Contrary to bandgap absorption but similarly to Co2+ absorption, Raman mea-
surements may give information concerning whether the phase transition is abrupt or
continuous. Figure 6.26(a) shows the pressure dependence of the intensity of the E2High
Raman mode. In order to determine the Raman intensity decrease due to the W-to-RS
structural phase transition onset, the intensity must be corrected for absorption effects
at the bandgap. Considering the Raman cross-section (eq. 2.10), and the fact that the
energy bandgap, 3.28 eV, separates from the excitation energy, 2.41 eV (514.5 nm), at a
rate of 17 meV/GPa, the compensated E2High intensity variation with pressure is depicted
in Fig. 6.26(b). A decrease of the intensity with increasing pressure is clearly detected
below 5 GPa, which points out the gradual character of the W-to-RS phase transition in
ZnO: 5%Co2+ nanoparticles.
Figure 6.26: Pressure dependence of the E2High experimental intensity (a) and normalized corrected
intensity (see text) (b) for ZnO: 5%Co2+ nanocrystals.
The pressure dependence of the E2High phonon energy up to the W→RS phase transition
is shown in Fig. 6.27 and has been fitted to a second-order polynomial [63]:
ω = ω(0) + aP + bP 2 (6.2)
with ω(0) = 438.4 cm−1 , a = 6.0 ± 0.2 cm−1 /GPa and b = 0.17 ± 0.03 cm−1 /GPa2 . The
experimental wavenumber at AP obtained in this work for the ZnO: 5%Co2+ nanocrystals,
438.4 cm−1 , is evidently red-shifted compared to the different bulk values, 444 cm−1 or
440 cm−1 given in the literature [27], [59]. This result is attributed to a relaxation of the
momentum conservation, and it confirms the confinement effect [61]. The linear frequency
dependence of E2High , a = 6.0 ± 0.2 cm−1 /GPa, is similar to the value reported for W-ZnO
single crystal, 5.2 cm−1 /GPa [59]. Above the transition pressure no active Raman modes
are observed. As it can be seen in Fig. 6.25, the studied nanoparticles remain in the
RS phase after the pressure down-stroke. Hence, Raman measurements also indicate the
metastability of the RS-ZnO: 5%Co2+ nanocrystals at ambient pressure.
Figure 6.27: Pressure dependence of the E2High optical phonon energy, Eph , for W-ZnO: 5%Co2+
nanoparticles.
6.3.3 Conclusions
The irreversible phase transition of W-ZnO and W-ZnO: 5%Co2+ colloidal nanocrystals
(4 nm in size) to the RS structure has been demonstrated by absorption and Raman
measurements. The ZnO nanoparticles direct absorption spectra have been measured as
a function of pressure showing a blue-shift (20 ± 2 meV/GPa) of the optical bandgap
with pressure up to 14 GPa, where the W-to-RS phase transition takes place. Absorp-
tion spectra recorded upon decreasing pressure indicate the metastability of the RS-ZnO
nanoparticles. Besides the bandgap absorption, the spectra of ZnO: 5%Co2+ nanocrys-
tals present a charge-transfer band and an absorption band due to Co2+ absorption in a
tetrahedral coordination. In this sample, the phase transition is gradually observed and
completed at around 14 GPa. This result is opposite to the abrupt transition detected
in ZnO: Co2+ single crystals and thin-films, and it is not fully understood yet. The ab-
sorption spectra during the down-stroke evidence the metastability of RS-ZnO: 5%Co2+
nanoparticles at RT and AP. The pressure dependence of the ZnO: 5%Co2+ E2High Raman
mode also shows that the W-to-RS phase transition is completed at ca. 15 GPa, and that
ZnO: 5%Co2+ nanocrystals remain in the RS structure at AP.
Bibliography
[1] Q. Jiang, C.C. Yang, and J.C. Li. Melting enthalpy depression of nanocrystals.
Mater. Lett., 56: 1019–1021, 2002.
[2] S. Xiao, W. Hu, and J. Yang. Melting Behaviors of Nanocrystalline Ag. J. Phys.
Chem. B, 109: 20339–20342, 2005.
[3] S.H. Tolbert and A.P. Alivisatos. High−Pressure Structural Transformations in Semi-
conductor Nanocrystals. Annu. Rev. Phys. Chem., 46: 595425, 1995.
[4] S.M. Clark, S.G. Prilliman, C.K. Erdonmez, and A.P. Alivisatos. Size dependence
of the pressure-induced γ to α structural phase transition in iron oxide nanocrystals.
Nanotech., 16: 2813–2818, 2005.
[5] S.H. Tolbert and A.P. Alivisatos. Size Dependence of a First Order Solid-Solid
Phase Transition: The Wurtzite to Rock-Salt Transformation in CdSe Nanocrystals.
Science, 265: 373–376, 1994.
[6] K. Jacobs, D. Zaziski, E.C. Scher, A.B. Herhold, and A.P. Alivisatos. Activation
Volumes for Solid-Solid Transformations in Nanocrystals. Science, 293: 1803–1806,
2001.
[8] A.H. Mueller, M.A. Petruska, M. Achermann, D.J. Werder, E.A. Akhadov, D.D.
Koleske, M.A. Hoffbauer, and V.I. Klimov. Multicolor Light-Emitting Diodes Based
on Semiconductor Nanocrystals Encapsulated in GaN Charge Injection Layers. Nano
Lett., 5: 10391044, 2005.
177
178 BIBLIOGRAPHY

2006.
[11] H. Ohno, D. Chiba, F. Matsukura, T. Omiya, E. Abe, T. Dietl, Y. Ohno, and

K. Ohtani. Electric-field control of ferromagnetism. Nature, 408: 944–946, 2000.
[12] K. Sato and H. Katayama-Yoshida. Electronic structure and ferromagnetism of

transition-metal-impurity-doped zinc oxide. Physica B, 308-310: 904–907, 2001.
[13] K. Ueda, H. Tabata, and T. Kawai. Magnetic and electric properties of transition-
metal-doped ZnO films. Appl. Phys. Lett., 79: 988–990, 2001.
[15] M.G. Bawendi, A.R. Kortan, M.L. Steigerwald, and L.E. Brus. X-ray structural
characterization of larger CdSe semiconductor clusters. J. Chem. Phys., 91: 7282–
7290, 1989.
[16] M. Cardona and G. Harbeke. Optical Properties and Band Structure of Wurtzite-
Type Crystals and Rutile. Phys. Rev., 137: A1467–A1476, 1965.
[17] K.J. Chang, S. Froyen, and M.L. Cohen. Electronic band structures for zinc-blende
and wurtzite CdS. Phys. Rev. B, 28: 4736–4743, 1983.
[18] H.S. Nalwa, editor. Handbook of Nanostructured Materials and Nanotechnology. Aca-
demic Press, New York, 2000.
[19] R.W.G. Wyckoff, editor. Crystal Structures, volume 2. John Wiley and Sons, New
York, London, 1963.
[20] Y.N. Xu and W.Y. Ching. Electronic, optical, and structural properties of some
wurtzite crystals. Phys. Rev. B, 48: 4335–4351, 1993.
BIBLIOGRAPHY 179
[21] C. Kumpf, R.B. Neder, F. Niederdraenk, P. Luczak, A. Stahl, M. Scheuermann,

S. Joshi, S.K. Kulkarni, C. Barglik-Chory, C. Heske, and E. Umbach. Structure de-
termination of CdS and ZnS nanoparticles: Direct modeling of synchrotron-radiation
diffraction data. J. Chem. Phys., 123: 224707–1–224707–6, 2005.
[22] A.L. Edwards and H.G. Drickamer. Effect of Pressure on the Absorption Edges of
Some III-V, II-VI, and I-VII Compounds. Phys. Rev., 122: 1149–1157, 1961.
[23] B. Batlogg, A. Jayaraman, J.E. Van Cleve, and R.G. Maines. Optical absorption,
resistivity, and phase transformation in CdS at high pressure. Phys. Rev. B, 27: 3920–
3923, 1983.
[24] M. Haase and A.P. Alivisatos. Arrested solid-solid phase transition in 4-nm-diameter
cadmium sulfide nanocrystals. J. Phys. Chem., 96: 6756–6762, 1992.
[25] A. San-Miguel. Nanomaterials under high-pressure. Chem. Soc. Rev., 35: 876–889,
2006.
[26] J. Gonzalez, F.V. Perez, E. Moya, and J.C. Chervin. Hydrostatic Pressure Depen-
dence of the Energy Gaps of CdTe in the Zinc-Blende and Rocksalt Phases. J. Phys.
Chem. Solids, 56: 335–340, 1995.
[27] C.A. Arguello, D.L. Rousseau, and S.P.S. Porto. First-Order Raman Effect in
Wurtzite-Type Crystals. Phys. Rev., 181: 1351–1363, 1969.
[28] U. Venkateswaran, M. Chandrasekhar, and H.R. Chandrasekhar. Luminescence and

Raman spectra of CdS under hydrostatic pressure. Phys. Rev. B, 30: 3316–3319,
1984.
[29] M. Abdulkbadar and B. Thomas. Study of Raman spectra of nanoparticles of CdS

and ZnS. Nanostruct. Mater., 5: 289–298, 1995.
[30] J.F. Scott and T.C. Damen. Raman scattering from surface modes of small CdS
crystallites. Optics Communications, 5: 410–412, 1972.
180 BIBLIOGRAPHY
[31] N.B. Owen, P.L. Smith, J.E. Martin, and A.J. Wright. X-ray Diffraction at ultra-high
pressures. J. Phys. Chem. Solids, 24: 1519–1524, 1963.
[32] H. Sowa. On the mechanism of the pressure-induced wurtzite- to NaCl-type phase

transition in CdS: an X-ray diffraction study. Solid State Sci., 7: 73–78, 2005.
[33] G.A. Samara and A.A. Giardini. Compressibility and Electrical Conductivity of
Cadmium Sulfide at High Pressures. Phys. Rev., 140: A388–A395, 1965.
[34] T. Suzuki, T. Yagi, S. Akimoto, T. Kawamura, S. Toyoda, and S. Endo. Compression

behavior of CdS and BP up to 68 GPa. J. Appl. Phys., 54: 748–751, 1983.
[35] J.Z. Jiang, J.S. Olsen, L. Gerward, and S. Mørup. Enhanced bulk modulus and
reduced transition pressure in γ−Fe2 O3 nanocrystals. Europhys. Lett., 44: 620–626,
1998.
[36] Q.F. Gu, G. Krauss, W. Steurer, F. Gramm, and A. Cervellino. Unexpected High
Stiffness of Ag and Au Nanoparticles. Phys. Rev. lett., 100: 045502, 2008.
[37] R. Heller, J. McGannon, and A. Weber. J. Appl. Phys., 21: 1283, 1950.
[38] C.H. Bates, W.B. White, and R. Roy. New High-Pressure Polymorph of Zinc Oxide.
Science, 137: 993, 1962.
[39] S. Desgreniers. High-density phases of ZnO: Structural and compressive parameters.

Phys. Rev. B, 58: 14102, 1998.
[40] F.J. Manjón, K. Syassen, and R. Lauck. Effect of Pressure on Phonon Modes in
Wurtzite Zinc Oxide. High Press. Res., 22: 299–304, 2002.
[41] D. Vogel, P. Krüger, and J. Pollmann. Ab initio electronic-structure calculations for

II-VI semiconductors using self-interaction-corrected pseudopotentials. Phys. Rev.
B, 52: R14316–R14319, 1995.
[42] A. Segura, J.A. Sans, F.J. Manjón, A. Munoz, and M.J. Herrera-Cabrera. Optical
properties and electronic structure of rock-salt ZnO under pressure. Appl. Phys.
Lett., 83: 278–280, 2003.
BIBLIOGRAPHY 181
[43] S.J. Gilliland. Structural, Optical and Magnetic Characterization of Pulsed Laser
Deposited Thin Films of Zn1−x Mx O (M=Mn, Fe, Ni, Cu) Transparent Magnetic
Alloys. PhD thesis, University of Valencia, 2008.
[44] G. Apai, J.F. Hamilton, J. Stohr, and A. Thompson. Extended X-Ray Absorp-
tion Fine Structure of Small Cu and Ni Clusters: Binding-Energy and Bond-Length
Changes with Cluster Size. Phys. Rev. Lett., 43: 165–169, 1979.
[47] J.A. Sans, A. Segura, F.J. Manjón, B. Marı́, A. Munoz, and M.J. Herrera-Cabrera.
Optical properties of wurtzite and rock-salt ZnO under pressure. Microelectronics
Journal, 36: 928–932, 2005.
[48] R.S. Kumar, A.L. Cornelius, and M.F. Nicol. Structure of nanocrystalline ZnO up
to 85 GPa. Curr. Appl. Phys., 7: 135–138, 2007.
[49] J.Z. Jiang, J.S. Olsen, L. Gerward, D. Frost, D. Rubie, and J. Peyronneau. Structural
stability in nanocrystalline ZnO. Europhys. Lett., 50: 48–53, 2000.
[50] E. Grzanka, S. Gierlotka, S. Stelmakh, B. Palosz, T. Strachowski, A. Swiderska-

Sroda, G. Kalisz, W. Lojkowski, and F. Porsch. Phase transition in nanocrystalline
ZnO. Z. Kristallogr. Suppl., 23: 337–342, 2006.
[51] W. Shan, W. Walukiewicz, J.W. Ager III, K.M. Yu, Y. Zhang, S.S. Mao, R. Kling,
C. Kirchner, and A. Waag. Pressure-dependent photoluminescence study of ZnO
nanowires. Appl. Phys. Lett., 86: 153117, 2005.
[52] F. Decremps, J. Pellicer-Porres, F. Datchi, J.P. Itié, A. Polian, F. Baudelet, and J.Z.
Jiang. Trapping of cubic ZnO nanocrystallites at ambient conditions. Appl. Phys.
Lett., 81: 4820–4822, 2002.
182 BIBLIOGRAPHY
[53] J.A. Sans, A. Segura, J.F. Sánchez-Royo, Ch. Ferrer-Roca, and E. Guillotel. Pressure
dependence of the optical properties of wurtzite and rock-salt Zn1−x Cox O thin films.
phys. stat. sol. (b), 244: 407–412, 2007.
[54] S.G. Gilliland, J.A. Sans, J.F. Sánchez-Royo, G. Almonacid, and A. Segura. Charge-
transfer absorption band in Zn1−x Mx O (M: Co, Mn) investigated by means of pho-
toconductivity, Ga doping, and optical measurements under pressure. Appl. Phys.
Lett., 96: 241902, 2010.
[55] P. Koidl. Optical absorption of Co2+ in ZnO. Phys. Rev. B, 15: 2493–2499, 1977.
[56] D.R. Stephens and H.G. Drickamer. Effect of Pressure on Tetrahedral Ni++ and
Co++ Complexes. J. Chem. Phys., 35: 429–435, 1961.
[57] J.F. Liu, Y. He, W. Chen, G.Q. Zhang, Y.W. Zeng, T. Kikegawa, and J.Z. Jiang.
Bulk Modulus and Structural Phase Transitions of Wurtzite CoO Nanocrystals. J.
Phys. Chem. C, 111: 2–5, 2007.
[58] M. Millot, J. González, I. Molina, B. Salas, Z. Golacki, J.M. Broto, H. Rakoto, and
M. Goiran. Raman spectroscopy and magnetic properties of bulk ZnO: Co single
crystal. J. Alloy Compd., 423: 224–227, 2006.
[59] F. Decremps, J. Pellicer-Porres, A.M. Saitta, J.C. Chervin, and A. Polian. High-
pressure Raman spectroscopy study of wurtzite ZnO. Phys. Rev. B, 65: 092101,
2002.
[60] J.M. Calleja and M. Cardona. Resonant Raman scattering in ZnO. Phys. Rev. B,
16: 3753–3761, 1977.
[61] J. Marquina, Ch. Power, and J. González. Raman scattering on ZnO nanocrystals.
Rev. Mex. Fis., 53: 170–173, 2007.
[62] X. Yan, Y. Gu, X. Zhang, Y. Huang, J. Qi, Y. Zhang, T. Fujita, and M. Chen. Doping
Effect on High-Pressure Structural Stability of ZnO Nanowires. J. Phys. Chem. C,
113: 1164–1167, 2009.
BIBLIOGRAPHY 183
[63] J. González, B.J. Fernández, J.M. Besson, M. Gauthier, and A. Polian. High-pressure
behaviour of Raman modes in CuGaS2 . Phys. Rev. B, 46: 15092–15101, 1992.
184 BIBLIOGRAPHY
Chapter 7
Conclusiones
Las conclusiones más importantes del trabajo desarrollado y presentado en esta tesis
pueden resumirse de la siguiente manera:
• Se han sintetizado nanopartı́culas de Y2 O3 : Er3+ , Yb3+ por dos métodos distintos;

en un molino de bolas planetario y mediante una reacción de combustión. En el caso
de las muestras preparadas en el molino, se alcanza un tamaño de partı́cula esta-
cionario de unos 25 nm, mientras que el tamaño estimado para las nanopartı́culas
obtenidas por combustión es de 10-15 nm. Sin embargo, es necesario calcinar las
muestras para reducir la contaminación en la superficie de las mismas, y el tamaño fi-
nal obtenido es de unos 50 nm. Todas las muestras se han protegido introduciéndolas
en capilares o recubriéndolas con SiO2 para el estudio de sus propiedades ópticas.
La presencia de la capa de SiO2 en la superficie podrı́a ser el paso inicial para la fun-
cionalización de las muestras nanocristalinas de Y2 O3 : Er3+ , Yb3+ . Se ha estudiado
la dependencia de las propiedades ópticas con el método de sı́ntesis o la concen-
tración de impurezas. La relación de intensidades de las emisiones roja y verde del
Er3+ puede modificarse cambiando la concentración de Yb3+ . La emisión de UC
es más intensa y presenta tiempos de vida más largos en el caso de las muestras
preparadas en el molino. Se ha identificado un aumento de la emisión roja cuando
se excita en el IR. Se ha establecido que el mecanismo principal responsable de la
luminiscencia de UC es un proceso GSA/ETU.
185
186
• Se han preparado muestras nanocristalinas de NaYF4 : 2%Er3+ , 20%Yb3+ en un

molino planetario de dos maneras diferentes, obteniendo en ambos casos un tamaño
de partı́cula de unos 30 nm. La muestra obtenida tras la molienda del polvo mi-
crocristalino (bulk ) durante 1 hora presenta la eficiencia de UC más alta pero sin
embargo es todavı́a mucho menos eficiente que el bulk. Esta disminución de la efi-
ciencia se ha relacionado con la contaminación de la superficie de los nanocristales
durante el proceso de molienda, la aparición de la fase cúbica α-NaYF4 , y el aumento
del cociente superficie/volumen en las nanopartı́culas.
• Se han preparado nanopartı́culas de GGG y YAG co-dopadas con Tb3+ -Yb3+ y

Eu3+ -Yb3+ mediante el método Pechini de sı́ntesis sol-gel, con un tamaño de partı́cula
de unos 30-40 nm. Se han estudiado e identificado las emisiones verde y azul del
Tb3+ ası́ como la luminiscencia roja del Eu3+ . Se ha observado emisión visible de
UC en ambos iones, Tb3+ y Eu3+ , tras la excitación en el IR a 10250 cm−1 . Tanto
los resultados experimentales como el modelo teórico de rate equations confirman
la sensitización cooperativa como el mecanismo responsable de la luminiscencia de
UC. Se ha comprobado que la presencia de impurezas no deseadas de Er3+ y su
emisión tras la excitación a 10250 cm−1 en las muestras co-dopadas con Eu3+ -Yb3+
no es relevante para la luminiscencia de UC del Eu3+ .
• Se ha llevado a cabo un estudio espectroscópico exhaustivo en muestras micro-

cristalinas de LMA impurificado con distintas concentraciones de Mn2+ e Yb3+ . El
sistema LMA cristaliza en la fase magnetoplumbita en la cual los iones de Mn2+ ocu-
pan sitios de coordinación tetrahédrica. Se ha observado luminiscencia de UC en el
Mn2+ ası́ como emisión de pares de Yb3+ en todas las muestras. Se ha demostrado
por espectroscopia resuelta en tiempo y experimentos en función de la concentración,
que la luminiscencia de UC en LMA: 1%Mn2+ , 1%Yb3+ y LMA: 2%Mn2+ , 5%Yb3+
tiene lugar mediante el mecanismo GSA/ESA, mientras que existe una contribución
GSA/ETU adicional en LMnA: 1%Yb3+ . El modelo teórico de rate equations con-
firma los mecanismos de UC propuestos considerando la formación de dı́meros Mn2+ -
Yb3+ . Se ha observado luminiscencia de UC en todas las muestras a temperatura
7. Conclusiones 187
ambiente e incluso por encima de 500 K en el caso de LMA: 2%Mn2+ , 5%Yb3+ ,

siendo esta emisión de UC un orden de magnitud más intensa. La luminiscencia de
UC en LMA: 2%Mn2+ , 5%Yb3+ es muy intensa hasta temperatura ambiente, pero a
temperaturas más elevadas tanto la intensidad como el tiempo de vida disminuyen
de la misma manera debido a procesos no radiativos. Por el contrario, en el caso de
LMnA: 1%Yb3+ se ha detectado una disminución de la intensidad de emisión por
encima de 100 K mientras que el tiempo de vida permanece constante hasta RT,
lo que se ha atribuido a una disminución de la excitación efectiva de los iones de
Mn2+ .
• Se ha estudiado la dependencia de las propiedades ópticas de nanopartı́culas de

GGG impurificadas con Cr3+ en función de la temperatura y la presión. Es bien
conocido que las técnicas de alta presión permiten modificar el acoplamiento de los
estados 4 T2 y 2 E del Cr3+ de manera continua. Al disminuir la temperatura se ha
observado una transformación del espectro de emisión pasando de una banda ancha
debida a la transición 4 T2 → 4 A2 , a lineas finas asociadas a la luminiscencia desde el
estado 2 E, acompañada de un aumento del tiempo de vida. Este comportamiento
en función de la temperatura es análogo a la dependencia con la presión observado
anteriormente en el bulk. Por el contrario, la influencia de la alta presión en la
luminiscencia para los nanocristales de GGG: Cr3+ es ligeramente distinta. Aunque
se observa un cambio de banda ancha a emisión tipo rubı́, la estructura fina no se
resuelve a RT ni siquiera a 20 GPa. Se ha modelizado la dependencia del tiempo de
vida con la presión y la temperatura teniendo en cuenta el acoplamiento espı́n-órbita
entre los estados 4 T2 y 2 E.
• Se han sintetizado nanopartı́culas de CdS con estructura ZB de unos 5 nm de

tamaño en un molino de bolas planetario. La transición de fase ZB → NaCl se ha
demostrado mediante medidas experimentales de absorción óptica, espectroscopia
Raman y difracción de rayos X bajo presión. El gap de energı́a directo en la fase ZB
aumenta con la presión siendo el coeficiente lineal 31 ± 4 meV/GPa, el cambio de
188
gap directo a indirecto tiene lugar por encima de 6 GPa, y en la fase NaCl la depen-
dencia del gap indirecto con la presión es de −35 ± 5 meV/GPa. La energı́a de los
modos Raman LO y 2LO aumenta con la presión, la desaparición de dichos modos
por encima de 6.0-6.5 GPa es también una evidencia clara de la transición de ZB a
NaCl en nanopartı́culas de CdS. Asimismo, las medidas Raman muestran que en la
descompresión la estructura ZB se recupera a presión ambiente. Los experimentos
de difracción de rayos X bajo presión también prueban que la transición de fase en
los nanocristales de CdS ocurre a partir de 5.4 GPa. Los valores del módulo de bulk
obtenidos para las estructuras ZB y NaCl son, respectivamente, B0 = 74 ± 2 GPa
y B0 = 107 ± 5 GPa. Todos los experimentos llevados a cabo demuestran que la
presión a la que se produce la transición es mucho mayor que en el caso del bulk.
• La transición de fase irreversible W → NaCl en nanopartı́culas de ZnO y ZnO:

5%Co2+ (4 nm de tamaño) se ha puesto de manifiesto en experimentos de ab-
sorción y Raman a alta presión. Los espectros de absorción en función de la presión
en nanocristales de ZnO muestran un desplazamiento del gap óptico hacia el azul
(20 ± 2 meV/GPa) hasta 14 GPa; produciéndose la transición de fase a esa presión.
Los espectros de absorción en el caso de nanocristales de ZnO: 5%Co2+ presentan,
además de la absorción del gap, una banda de transferencia de carga ası́ como una
banda de absorción asociada al Co2+ en coordinación tetrahédrica. En la muestra
impurificada la variación de la absorción del Co2+ indica que la transición de fase
tiene lugar de forma progresiva, empezando en torno a 4-5 GPa, hasta completarse
a 14 GPa. Este resultado es opuesto a la transición brusca previamente descrita
en bulk y láminas delgadas de ZnO: Co2+ . Los espectros de absorción al relajar
la presión indican la metaestabilidad de la fase NaCl para las nanopartı́culas en
condiciones ambiente. La dependencia con la presión de la energı́a del modo Raman
E2High también demuestra que la transición de fase W → NaCl se completa a 15
GPa, y que los nanocristales de ZnO: 5%Co2+ permanecen en la estructura NaCl a
presión ambiente.
List of Publications
1. R. Martı́n-Rodrı́guez, R. Valiente, C. Pesquera, F. González, C. Blanco, V. Potin

and M.C. Marco de Lucas. Optical properties of nanocrystalline-coated
Y2 O3 :Er3+ , Yb3+ obtained by mechano-chemical and combustion syn-
thesis. J. Lumin., 129, 1109-1114 (2009).
2. R. Martı́n-Rodrı́guez, R. Valiente, S. Polizzi, M. Bettinelli, A. Speghini and F.

Piccinelli. Upconversion luminescence in nanocrystals of Gd3 Ga5 O12
and Y3 Al5 O12 doped with Tb3+ -Yb3+ and Eu3+ -Yb3+ . J. Phys. Chem. C,
113, 12195-12200 (2009).
3. R. Martı́n-Rodrı́guez, R. Valiente and M. Bettinelli. Room-temperature

green upconversion luminescence in LaMgAl11 O19 : Mn2+ , Yb3+ upon
infrared excitation. Appl. Phys. Lett., 95, 091913 1-3 (2009).
4. R. Valiente, F. Rodrı́guez, J. González, H.U. Güdel, R. Martı́n-Rodrı́guez, L.

Nataf, M.N. Sanz-Ortiz and K. Krämer. High pressure optical spectroscopy
of Ce3+ -doped Cs2 NaLuCl6 . Chem. Phys. Lett., 481, 149-151 (2009).
5. R. Martı́n-Rodrı́guez, R. Valiente, F. Rodrı́guez and J. González. Optical

energy gap on zinc-blende CdS nanoparticles under high pressure.
High. Press. Res., 29(4), 482-487 (2009).
6. R. Martı́n-Rodrı́guez, R. Valiente, F. Rodrı́guez, F. Piccinelli, A. Speghini and

M. Bettinelli. Temperature dependence and temporal dynamics of Mn2+
upconversion luminescence sensitized by Yb3+ in codoped LaMgAl11 O19 .
Phys. Rev. B, 82, 075117 1-7 (2010).
189
190
7. C. Renero-Lecuna, R. Valiente, R. Martı́n-Rodrı́guez, J. González, F. Rodrı́guez,

K. Krämer and H.U. Güdel. Multi-site selective spectroscopy and high
pressure upconversion in NaYF4 : Er3+ , Yb3+ . In preparation.
8. R. Martı́n-Rodrı́guez, R. Valiente, F. Rodrı́guez and M. Bettinelli. Temper-

ature and pressure dependence of the optical properties of Cr3+ -
doped GGG nanoparticles. In preparation.
9. R. Martı́n-Rodrı́guez, J. González, R. Valiente, F. Aguado, D. Santamarı́a and

F. Rodrı́guez. X-ray diffraction, optical absorption and Raman mea-
surements of CdS under high pressure. In preparation.
10. R. Martı́n-Rodrı́guez, R. Valiente, J. González, G. Almonacid, A. Segura, F.

Aguado, F. Rodrı́guez, M.A. White and D.R. Gamelin. Metastability of Wurtzite
and Rock-salt phases in Zn1−x Cox O nanoparticles investigated by op-
tical absorption and Raman spectroscopy under high pressure. In
preparation.

SYNTHESIS, STRUCTURAL CHARACTERIZATION AND SPECTROSCOPIC STUDY OF NANOCRISTALLINE AND MICROCRYSTALLINE MATERIALSmicrocristales

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SYNTHESIS, STRUCTURAL CHARACTERIZATION AND SPECTROSCOPIC STUDY OF NANOCRISTALLINE AND MICROCRYSTALLINE MATERIALSmicrocristales

Uploaded by

Copyright:

Available Formats

UNIVERSIDAD DE CANTABRIA

Departamento de Fı́sica Aplicada

SYNTHESIS, STRUCTURAL CHARACTERIZATION AND

Rosa Martı́n Rodrı́guez

Santander, Noviembre de 2010

Que el trabajo que se presenta en esta memoria, titulado SÍNTESIS, CARACTERI-

Santander, Noviembre de 2010

Fdo.: Rafael Valiente Barroso

List of Abbreviations vii

5 Insulating materials. Luminescent properties 73

6 Semiconductor nanoparticles. Optical spectroscopy at high pressure 149

List of Publications 189

CCD Charge-coupled device

DAC Diamond anvil cell

DMSO Dimethyl sulfoxide

DMSs Diluted magnetic semiconductors

ESA Excited-state absorption

ESCO Excited-state crossover

ETU Energy transfer upconversion

G/N Glycine-to-nitrates ratio

GGG Gd3 Ga5 O12

GSA Ground-state absorption

LMA LaMgAl11 O19

LMnA LaMnAl11 O19

MIR Medium infrared

NIR Near infrared

OPO Optical parametric oscillator

PEG Polyethylene glycol

QDs Quantum dots

TEM Transmission electron microscopy

TEOS Tetraethyl orthosilicate

TOPO Trioctylphosphine oxide

XRD X-ray diffraction

YAG Y3 Al5 O12

ZPL Zero phonon line

La mayorı́a de los estudios de UC involucran combinaciones de iones de tierras raras,

absorción de estado excitado (GSA/ESA) y transferencia de energı́a (GSA/ETU). Por

Por último, se han estudiado las propiedades ópticas y estructurales de nanopartı́culas

X-ray diffraction (XRD).

2.2 Interest of nanoparticles

Moving to our research field, there is a considerable interest in the development of

2.3 Insulating materials. Optical properties

2.3.2 Ions in a static crystalline environment

2.3.3 The configurational coordinate diagram

ESS = 2S~ω (2.4)

show strong coupling giving rise to broad bands in their spectra.

2.3.4 Light absorption and emission processes

Figure 2.2: Classical interpretation of the activation energy, Ea , in non-radiative transitions.

2.3.5 Non-radiative transitions

Contrary to RE ions, in which luminescence is a quite common phenomenon even from

• Multiphonon relaxation: In these processes, the electronic excitation is totally or

2.4 Raman Spectroscopy

hv|Hint |cihc|Hep |cihc|Hint |vi

The intensity of Raman scattering is defined by I = P α2 ω04 , where P is the laser

2.6 Rare-earth ions

increase for larger atomic number.

2.7 Transition metal ions

field (Fig. 2.5(b)). Tanabe-Sugano diagrams allow to predict if an optical band of a TM

2.8 Semiconductor nanostructures

2.9 Interband absorption in semiconductors

2.9.2 Band edge absorption in direct gap semiconductors

g(~ω) = 0 For ~ω < Eg (2.18)

α(~ω) = 0 For ~ω < Eg (2.19)

α(~ω) ∝ (~ω − Eg )1/2 For ~ω ≥ Eg