Luminescent properties of semiconductor nanocrystals

Citation for published version (APA):Chin, P. T. K. (2008). Luminescent properties of semiconductor nanocrystals. Technische UniversiteitEindhoven. https://doi.org/10.6100/IR638886

DOI:10.6100/IR638886

Document status and date:Published: 01/01/2008

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

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

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

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

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:openaccess@tue.nlproviding details and we will investigate your claim.

Download date: 27. Jan. 2021

Luminescent Properties of Semiconductor Nanocrystals

Luminescent Properties of Semiconductor Nanocrystals

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de Rector Magnificus,

prof.dr.ir. C.J. van Duijn, voor een commissie aangewezen door het College voor Promoties in het openbaar te verdedigen op

woensdag 26 november 2008 om 16.00 uur

door

Patrick Ted-Khong Chin

geboren te Voorschoten

Dit proefschrift is goedgekeurd door de promotor: prof.dr.ir. R.A.J. Janssen This research has been financially supported by the Dutch Government trough the NanoNed and by the Interreg program OLED+ Omslagontwerp: Patrick Chin Druk: Gildeprint, Enschede A catalogue record is available from the Eindhoven University of Technology Library ISBN: 978-90-386-1455-7

Contents

List of Abbreviations Chapter 1 Introduction 1 1.1 History and early observations 2 1.2 Quantum size effects in semiconductors of finite size 3 1.3 Shape and properties 5

1.3.1 From dots to anisotropic structures 5 1.3.2 Shape control of colloidal semiconductor nanocrystals 7

1.4 Semiconductor heterostructures 9 1.4.1 Introduction to nanocrystal heterostructures 9 1.4.2 Type II heterostructures 10 1.4.3 Heterostructures with different crystal structures 11

1.5 Doped nanocrystals 12 1.6 Semiconductor nanocrystal applications 13

1.6.1 Introduction to applications 13 1.6.2 Hybrid organic and NC-LEDs 14

1.7 Aim of the thesis 15 References 16 Chapter 2 Energy Transfer in Hybrid Quantum Dot LEDs 19 2.1 Introduction 20 2.2 Experimental 21 2.3 Results and discussion 22

2.3.1 Energy transfer 22 2.3.2 Electroluminescence 27

2.4 Conclusion 31 References 32 Chapter 3 Highly Luminescent CdTe/CdSe Colloidal

Heteronanocrystals with Temperature Dependent Emission Color

35

3.1 Introduction 36 3.2 Experimental 37 3.3 Results and discussion 39

3.3.1 CdTe/CdSe heteronanocrystal growth 39 3.3.2 Optical properties: absorption, photoluminescence,

quantum yields and exciton lifetimes 41

3.3.3 Temperature dependence of the optical properties 45 3.4 Conclusion 49 References 50 Chapter 4 Cluster Synthesis of Branched CdTe Nanocrystals for

Light-Emitting Diodes 53

4.1 Introduction 54

4.2 Experimental 55 4.3 Results and discussion 57

4.3.1 Synthesis and characterization 57 4.3.2 Electroluminescence 64

4.4 Conclusion 67 References 68 Chapter 5 Polarized Light Emitting Quantum Rod Diodes 71 5.1 Introduction 72 5.2 Experimental 73 5.3 Results and discussion 74

5.3.1 Polarized Photoluminescence 74 5.3.2 Polarized Electroluminescence 76

5.4 Conclusion 79 References 79 Chapter 6 Energy Transfer and Polarized Emission in Cadmium

Selenide Nanocrystal Solids with Mixed Dimensionality 81

6.1 Introduction 82 6.2 Experimental 83 6.3 Results and discussion 85 6.4 Conclusion 93 References 94 Chapter 7 Highly Luminescent Ultra Thin Mn Doped ZnSe

Nanowires 97

7.1 Introduction 98 7.2 Experimental 98 7.3 Results and discussion 101 7.4 Conclusion 109 References 109 Summary 113 Samenvatting 117 List of publications 121 Curriculum Vitae 123 Dankwoord 125

Abbreviations Abs absorption a.u. arbitrary unit CB conduction band DDA dodecylamine EL electroluminescence ETL electron transport layer FRET fluorescence resonance energy transfer FWHM full width at half maximum HBL hole blocking layer HDA hexadecylamine HOMO highest occupied molecular orbital HPA hexylphosphonic acid HRTEM high resolution transmission electron microscopy HTL hole transport layer IR infrared ITO indium tin oxide LCAO linear combination of atomic orbitals LED light-emitting diode LUMO lowest unoccupied molecular orbital NC nanocrystal NR nanorod ODA octadecylamine PL photoluminescence PEDOT poly(3,4-ethylenedioxythiophene) PSS poly(styrenesulfonate) PSF poly(2,7-spirofluorene) PVK polyvinylcarbazole QD quantum dot QR quantum rod QY quantum yield TCSPC time correlated single photon counting TEM transmission electron microscopy TGA thermal gravimetric analysis TOP trioctylphosphine TOPO trioctylphophine oxide TDPA tetradecylphosphonic acid TPBI 1,3,5-tris(N-phenylbenzimidazol-2-yl)benzene UV-Vis ultraviolet visible VB valence band W wurtzite XRD X-ray diffraction ZB zinc blende

1

Chapter 1

Introduction

Summary

This chapter gives an introduction to the history and early observations of the size related physical properties of nanosized semiconductor materials. Nanosized materials show fascinating and unique differences in optical and electronic properties with respect to bulk materials. The distinct physical and chemical features of these nanomaterials make them an exciting and attractive class of novel materials with an enormous potential for various applications. The size related properties can be rationalized quantum mechanically using the concept of size quantization. In the nanometer regime, opto-electronic properties also depend on shape and can be controlled by the number of dimensions in which size is confined. The synthetic strategies towards nanocrystalline semiconductors with controlled size and shape are outlined. Opto-electronic properties also depend on the composition and in this respect nanocrystalline heterostructures are of interest. Heterostructures of different epitaxially grown crystalline materials allow selective carrier confinement and further control over both the emissive and electronic properties. The incorporation of atomic impurities is an alternative way to modify the physical properties of a semiconductor. Such doping can also be applied in nanosized semiconductors. This enables the formation of materials where the properties are determined by both size effects and atomic band transitions of the dopants.. The applications discussed in this work relate to the luminescent properties as active component in light-emitting diodes and nanophosphors.

2 Chapter 1

1.1 History and Early Observations Nanostructures with their dimensions between 1 to 100 nm attract an enormous attention in the last two decades. The development of the modern micro (electronic) integrated circuits stimulated an extensive research effort to create smaller structures in order to reach higher performances, less power consumption, and lower costs. The scaling down of bulk metals and semiconductors to the nanometer regime revealed several exciting phenomena, such as size-dependent excitation, quantisized conductance, and metal to semiconductor to insulator transitions.1 Modern physics and mathematics make it possible to study, simulate, and explain these size and shape dependent properties. The application of nanosized materials is however much older than today’s science, and dates back to ancient Egypt and Roman times. In the ancient times metal nanoparticles were formed in molten glass, and used to made stained glass objects. Such a magnificent example of ancient glass is the Lycurgus Cup2 (AD fourth century), illustrating myth of King Lycurgus (Figure 1). The dispersed gold nanoparticles in the glass make the glass appears green, when viewed in reflecting daylight. When the cup is illuminated from the inside it appears red by the transmitted light.

Figure 1. The Lycurgus Cup (British Museum).

The first study on the size dependence of the physical properties of metals was reported by Faraday 1856.3 Faraday observed that the electronic structure of a metal can become size dependent below a certain size. The size dependent phenomena were also observed in 1960’s for semiconductors,4,5 where for colloidal dispersions of AgBr and AgI a shorter absorbance wavelength was observed compared to macroscopic material.5 The study of layered MoS2 and quantum wells revealed spatial dependent optical properties for quantum wells with different layer thickness.6 Evans and Young were among the first who related these findings to size quantization of the electronic structure of a semiconductor.6

(a) (b)

Introduction 3

It took however until the 1980’s when the first theoretical explanation was proposed for colloidal spherical nanocrystals (NCs) by Brus.7 Together with advances in the synthetic procedures, 8,9 this lead to a rapid increase in research in the field of nanosized materials. A breakthrough in the synthesis of high quality monodisperse semiconductor NCs or quantum dots (QDs) was made by the work of Murray, Norris, and Bawendi in 1993.10 They separated the initial nucleation from the particle growth, by rapid injection of reagents into a hot coordinating solvent. The sudden increase in precursor concentration above the nucleation threshold at sufficient high temperature triggers a short burst of nucleation, leading to a rapid decrease in precursor concentration below the nucleation point. At this point, no new particles are formed and the growth proceeds by the consumption of monomers from solution by the QD nuclei. This “hot injection method” enabled the creation various types (CdSe, CdTe, CdS, PbSe, and ZnSe) of monodisperse (<10% size distribution) and high quality NCs. In recent years, various alternative methods have been developed resulting in high quality monodisperse NCs in both hydrophobic6,8 and aqueous environments.11,12,13 The work presented in this thesis is mainly based on the hot injection method6,14 and on the formation of colloidal QDs from temperature initiated growth using preformed atomic clusters.15 Colloidal semiconductor NCs with various shapes and properties, are prepared, studied, and used as a novel class of luminescent materials with distinct optical properties.

Figure 2: Exciton emission of a CdSe QDs dispersions in chloroform with decreasing size, under UV illumination.

1.2 Quantum Size Effects in Semiconductors of Finite Size The fascinating optical changes observed by Berry 4,5 and Brus7 for the reduced sized semiconductor NCs in colloidal dispersions can be related to an increase in band gap with decreasing particle size. When an electron in a semiconductor is promoted from the

4 Chapter 1

valance band to the conduction band through excitation, a hole in the valance band is created. This positive hole will form a bound state with the excited negative electron by Coulomb interaction. Such bound electron hole pair is often referred to as an (Mott-Wannier) exciton, and can be described in a similar way to the hydrogen-like bound state between the proton and the electron of the hydrogen atom. The spatial occupation of an exciton can be expressed in terms of an exciton Bohr radius (ab):

⎟⎟⎠

⎞⎜⎜⎝

⎛+= ∞

**20

20 114

heb mmem

aηεπε

(1.1)

Where ε∞ is the high frequency relative dielectric constant of the medium, me* and mh

* are the effective masses of the electron and the hole (in units of, the mass of an electron at rest m0). The elementary charge is represented by e, the vacuum permittivity by ε0, and ħ represents the Planck constant. The effective mass is defined as the mass that a particle seems to have in the classical model of transport in a crystal. The Bohr radius (ab) found for common semiconductors (CdSe, CdS, CdTe) is however much larger than for the hydrogen atom (a0). This is a consequence of the effective masses which are smaller than the mass of an electron at rest m0, and ε∞ is much larger than 1. When the particle size approaches that of the exciton Bohr radius, the exciton wave function becomes confined by the spatial limitations of the crystal. The potential barrier at the crystal surface forces the exciton wave function to go to zero at the crystal surface, confining the exciton wave function in the crystal. This will result in an increase in exciton energy with decreasing crystal size, corresponding to a blue shift in both exciton absorbance and emission (Figure 2). The size related optical and electronic phenomena are also known as “quantum size effect”. The increase of the band gap with decreasing size caused by the confinement of an exciton in a finite sized crystal, can be described using the “particle in a box” model. A solution for this model is presented by Brus equation7 for a spherical particle:

termssmallerR

emmRm

EEhe

g +−⎟⎟⎠

⎞⎜⎜⎝

⎛++=

∞επεπ

0

2

**20

22

4786.111

2η

(1.2)

The term Eg is the bulk semiconductor band gap, the second term is the solution of the Schrödinger equation for a particle in a spherical potential well, with a 1/R2 dependence where R represents the crystal radius. The third term describes the decrease in energy as result of the free electron and hole Coulomb attraction.

Introduction 5

A second effect which is observed as result of the decrease in crystal size is the appearance of discrete energy levels at the band edges. The formation of discrete energy levels can be clearly observed in absorbance spectra of monodisperse colloidal QDs with decreasing size (Figure 3a). Figure 3a shows the presence of clear an excitonic absorption peak, followed by a second and third absorption feature as result of the formation of discrete transitions above the band gap. This effect can be explained by the fact that QDs are an intermediate state between an atomic cluster and a bulk crystal. Semiconductors show a low density of states at the band edges and a quasi-continuum in band structure above and below the band gap (Figure 3b). As consequence of the limited number of atoms in a QD (100-10000) the density of states shows a molecular to atomic like behavior at the band edges. The discrete energy levels observed at the band edges become more pronounced with decreasing particle size as the separation between the levels increases.

350 400 450 500 550 600 650 7000.00

0.05

0.10

0.15

0.20

0.25(a)

Nor

mal

ized

Abs

orpt

ion

Wavelength (nm) Figure 3. (a) Normalized absorption spectra of HDA/TOPO capped CdSe QDs of different sizes in chloroform. (b) Schematic representation of the band structure of a bulk semiconductor crystal, a QD, and the atomic energy levels.

1.3 Shapes and Properties 1.3.1 From Dots to Anisotropic Structures The study on quantum size confinement which started in the 1960’s on MoS2 layers6 was followed by various studies on quantum wells16 where spatial confinement is in one direction only, i.e. perpendicular to the plane. In the 1980’s this was extended in a “zero” dimensional system by the creation of spherical QDs.7,8,9 In a spherical QD, the

(b)

6 Chapter 1

wave function is confined in all three principal directions and, hence, extended in zero dimensions (0 dimensional, 0D). The enormous progress in preparing monodisperse CdSe QDs opened the opportunity to extensively study the size related properties of these 0D systems as discussed in the previous paragraph. Recent approaches to tune the shape of colloidal NCs lead to the formation of nanorods, nanowires, and branched structures such as tetrapods and multipods. This opened an entire new area and enabled examining the behavior of the size effects for both the optical and electronic structure in 0, 1, and 2 dimensional extended semiconductors. Theoretical predictions about the quantum size confinement can be easily made using the effective mass approximation and the particle in a box model for 0D, 1D, and 2D systems. This shows the possibility to create novel classes of materials with shape dependent optical properties. Ignoring the distinct electronic structure of a material and using a simplified model, Figure 4 shows the evolution of the density of states from continuous levels of the 3D bulk material to discrete states in a 0D system

Figure 4. Simplified representation of the density of states in 3D, 2D, 1D and 0D semiconductors. The 1D quantum rods (QRs) are particular interesting because they show significant differences compared to spherical QDs (Figure 5). The QRs show a strong (up to 87%) linearly polarized photoluminescence (PL)17 and an increase of the global Stokes shift17,18 together with a faster carrier relaxation.19 Both experimental results and empirical pseudopotential calculations showed a crossover from non polarized emission for spherical QDs20 to linear polarization for elongated QRs.17,20 An alternative explanation for the linear polarization in QRs was provided by Shabaev and Efros21 who described the energy spectra and polarization properties of 1D excitons in QRs. The polarization of the luminescence is ascribed to the fine structure of the ground exciton state, which is split by the electron-hole exchange interaction that mixes different electron (sz) and hole (jz) spin

Introduction 7

states. For CdSe semiconductor QRs, a thermally populated Fz = sz + jz = 0 state (“0b”), causes the emission parallel to the rod to be more intense than in the perpendicular direction. Moreover, the emission from the “0b” state is much stronger than from the second degenerate optically active state “1±” state, as a result of a strong reduction of the electric field of a photon, if the field is perpendicular to the QR axis, and is almost unchanged if the field is parallel to the QR axis.21 These distinct shape dependent optical properties of the anisotropic NCs are employed in this thesis in Chapter 5 and 6 for application of CdSe/CdS nanorods as a novel luminescent emitter in both a linearly polarized QR-LEDs and in combination with QDs as a nanocrystalline emitter that absorbs non-polarized light and emits polarized light. 1.3.2 Shape Control of Colloidal Semiconductor Nanocrystals The popular hot injection method, 10 where fast nucleation is followed by a controlled and slower growth, leads to the formation of spherical or nearly spherical particles. A spherical shape represents the thermodynamic lowest-energy shape for materials with a relatively isotropic underlying crystal structure.22 Materials with a relatively anisotropic underlying crystal structure will often also form nearly spherical nanoparticles, as result of the importance of the surface in the nanosize regime, because the surface energy is minimized in spherical particles compared to anisotropic structures. Crystal shape and growth direction can be largely controlled by influencing both kinetic and thermodynamic parameters. The precursor concentration in the reaction mixture is an important factor to control the kinetic growth conditions. High precursor concentrations will promote a fast kinetic growth, leading to a more anisotropic growth, especially in systems where the underlying crystal structure is anisotropic. A second important parameter is the ability of specific surfactant molecules to bind with different affinities to certain NC facets, controlling the relative reactivity of the surface facet.23,24 Surface energy and binding energy calculations of the various facets of CdSe show a difference in reactivity and affinity of the different facets of wurtzite CdSe towards monomers and surfactants.25 The surface binding of monomers and surfactants combined with factors as temperature and precursor concentration, play a key role in shape control of semiconductor NCs.

8 Chapter 1

20 nm20 nm

Figure 5. Transmission electron micrographs of (a) spherical CdSe QDs, (b) elongated CdSe QRs, and (c) high resolution image showing the wurtzite lattice fringes. Another approach to shape control discussed in this thesis, is so called oriented attachment. Oriented attachment is described as the phenomenon to create nanowires by connecting existing individual NCs, so that they share a common crystallographic orientation.26,27,28 This phenomenon is especially relevant for nanosized particles, where bonding between particles reduces the overall energy by decreasing surface energy caused by unsatisfied bonds.26 The self-assembly of individual nanoparticles has been studied for various materials: Ag,29 CdSe,27 CdTe,30 PbSe,28 and ZnO.31 The difference in crystal facet reactivity and the anisotropy of the crystal structure was found to be the driving force for oriented self-assembly in previous studies.27-31 Dipolar interactions caused by the crystal anisotropy are suggested to be the main driving force to orient rock salt PbSe NCs into wires.28 Dipolar interactions can originate from opposite terminated crystal facets, inducing a dipole moment in the particles. Dipole moments can be created even in highly symmetric crystal structures such as rock salt PbSe. Permanent dipole moments have been also reported in centrosymmetric zinc blende ZnSe NCs.32

Figure 6. Transmission electron micrograph of single crystalline ZnSe nanowires formed by oriented attachment. A small amount of non attached spherical particles can also be observed, their presence is discussed in Chapter 7 (the scale bar is 50 nm).

50 nm50 nm

Introduction 9

In anisotropic growth, selective binding of surfactants, reaction temperature, and precursor concentrations are important parameters that control the dipole arrangement on the crystal. Oriented attachment opens the opportunity to create anisotropic structures with materials showing natural isotropic crystal structure (Figure 6). This opportunity is exploited in this research to create zinc blende ZnSe nanowires doped with manganese. Effective doping of ZnSe is preferred when both ZnSe and MnSe have both the zinc blende crystal structure,33 which will naturally result in spherical doped particles. Chapter 7 discusses the creation of doped ZnSe nanowires further, revealing a novel class of anisotropic materials with distinct optical properties.

1.4 Semiconductor Heterostructures 1.4.1 Introduction to Nanocrystal Hetrostructures The NCs discussed so far consist of a single semiconductor material. Excitations in these NCs can be regarded as “a particle in a box” system,7 where the surface of the particle acts as the wall of the box. For a perfect isolated particle these walls will form an infinite potential barrier. In reality this is never the case and the height of the potential barriers is determined by the crystal surface properties and the surrounding medium. The atoms at the crystal surface can give rise to energy states that are different with respect to the bulk. Surface states can act as centers for non-radiative decay34 or lead to surface trap emission,35 resulting in a decrease in luminescence efficiency. The surface of these colloidal particles is usually covered with an organic coating consisting of surface bound surfactants and ligands. These surfactants and ligands are necessary during synthesis to control the growth, solubility, crystal morphology, and passivation of crystal defects and surface states.36 Controlling both the organic and inorganic surface chemistry is therefore very important to control both the physical and chemical properties of these colloidal NCs. The tunability of the surface related properties make these colloidal NCs unique compared the epitaxial formed NCs. Inorganic surface modification has shown to be a successful effective method for passivation of the crystal surface defect states. Overcoating monodisperse CdSe QDs with epitaxial layers of ZnS37,38 or CdS36,39 can now be routinely performed. This results in a dramatic improvement of the PL quantum yield (QY) (> 70%) compared to NCs that are solely capped by organic surfactants or ligands. The enhanced PL QY results from the decrease in dangling bonds and surface defects, which leads to an improved carrier and exciton confinement in the core of the QDs.14,38,40 The improved carrier and exciton

10 Chapter 1

confinement in QD core occurs when the band energies of the shell are such that valance band is lower and the conduction band is higher than those of the core material (Figure 7). This type of core-shell QD is also known as type-I QD.38,41

To avoid homogeneous nucleation and growth of shell material, the shell precursor concentrations should be below the nucleation threshold and the temperature need to be low during the overcoating reaction. The low precursor concentration supports undersatured-solution conditions leading to heterogeneous growth, which is often achieved by slow dropwise addition of precursors at moderate temperatures. The crystal structure and lattice constants of both the core and shell material are important factors for successful overcoating. Small spherical QDs like CdSe can be successfully overcoated with a ZnS shell despite a 12% lattice mismatch. The reduced facet length and high curvature in such spherical core-shell system allows an epitaxial growth despite the large lattice mismatch. Such a large lattice mismatch however will cause significant linear lattice strain when overcoating a rod shaped NCs, like CdSe with ZnS, and will result in lattice faults and defects with increasing shell thickness. As a result of the lattice mismatch between core and the shell material a decrease in PL QY is often also observed for spherical particles with thick shells (>3 monolayers).41

1.4.2 Type-II Heterostructures Another important class of (core-shell) heterostructures is formed by type-II QDs.42,43,44 Type-II QDs show several exciting properties that differ substantially from the type-I QDs. Spatial separation and selective confinement of excited charge carriers will occur in these QDs.43,44 This effect is achieved by overcoating the core with a material of which the band levels are shifted compared to the core material (Figure 7). Type-II QDs can give rise to selective carrier confinement. This occurs, for example, in CdTe/CdSe core-shell particles were the excited electron is mainly located in the CdSe shell and hole is confined in CdTe core. As a result of the selective carrier localization, the PL of the QDs is determined by the band offsets of two materials. The recombination of the charge carriers will occur over the interface between the two materials. This results often in a strongly red shifted PL together with long radiative lifetime with respect to the emission of the core QD itself. 43,44

Introduction 11

Figure 7. (a) Schematic representation on a core-shell QD with the organic ligands. (b) The energy diagrams for a type-I QD with a confined electron hole pair in the core. (c) The energy diagram for a type-II QD, charge separation occurs at the type-II interface leaving the hole confined in the core and the electron in the shell.

1.4.3 Heterostructures with Different Crystal Structures Differences in crystal structure between core and shell material can lead to facet selective growth. Facet selective particle growth has already been discussed in paragraph 1.3 where a growth direction on a crystal facet could be influenced by selective termination of a facet with strong binding surfactants, leading to shape control during particle growth. Another approach to create different shapes with either type-I or type-II behavior can be accomplished by growing two materials with different crystal structures onto each other. Different shapes varying from nanobarbells45 to tetrapods46,47 and branched structures48 have been shown in previous studies. Depending on the their synthetic route, for example CdSe and CdTe can both be created with zinc blende24,49,50 (cubic) or with a wurtzite10, 14,24 (hexagonal) crystal structure. Shape control of heterostructures was recently demonstrated by Alivisatos et al.46,47 They first created zinc blende CdSe QDs followed by CdS or CdSe “shell growth” in a reaction mixture with phosphonic acids present that favor wurtzite particle growth. This results in selective growth of wurtzite CdX (X= S or Se) legs on the spherical zinc blende CdSe core, resulting in the formation of tetrapods. This effect can be explained by the fact that the wurtzite crystal structure of CdS and CdSe can easily nucleate on the {111} facets of zinc blende CdSe QDs. The zinc blende {111} facet of CdSe is structurally similar to the {001} facets of wurtzite CdSe and CdS.51,52

(a) (b) (c)

12 Chapter 1

Figure 8. Transmission electron micrographs of (a) CdTe/CdSe QDs, (b) CdTe/CdS (c) CdTe/CdSe. In Chapters 3 and 4, this concept is further extended, showing the growth of wurtzite CdSe or CdS on zinc blende CdTe QDs, resulting in the creation of branched heterostructured NCs (Figure 8) revealing eventually a strong type-II character.

1.5 Doped Nanocrystals The incorporation of atomic impurities in semiconductors is a common method to modify and tailor the electrical and optical properties of semiconductors. This technique is also known as doping. Selective doping of semiconductors, such as silicon and germanium, enables the creation of p-n diodes and transistors which eventually lead to integrated circuits and microelectronics that we use in every day live. The doping of semiconductor NCs can also result in drastic changes of both optical and electronic properties. The ability to use atomic impurities in NCs therefore extends the already exciting size-depend properties as shown previously for doped NCs like CdSe, CdS, ZnSe, and ZnS doped with Cu, Mn and more exotic dopants.53,54 These dopants often function as emissive traps, with trap energy levels between the valence and conduction band of the host crystal. The creation of highly luminescent and stable doped NCs turns out to be a difficult task. In previous work the luminescence efficiency often does not exceed a QY of 20%,55 which is rather poor compared to the 70% routinely achieved for common core-shell QDs. Only recently, work by Peng et al. showed the creation of highly efficient ZnSe:Mn QDs with QYs up to 70%.56 The commonly used transition metal Cu and Mn dopants, are substitutional as they are located in the host lattice replacing the metal ion of the host. Recent work by Norris et al.33 shows that the match between the crystal structure and lattice constants of the dopant and the host material is crucial for effective incorporation of the dopants. The system discussed in this thesis is based on ZnSe doped with Mn. The Mn2+ ion shows visible luminescence due to

Introduction 13

a strong sensitivity to the crystal field of the host ZnSe lattice. The emission arises from the transition between the d-orbitals of the manganese atom (4T1→6A1), and can be shifted by influencing the crystal field splitting. Excitation of the Mn2+ ion is expected to be the result of carrier or exciton trapping. The band alignment of the valence and conduction band of ZnSe are such that the atomic Mn trap levels are in a type-I arrangement, and therefore function as a (emissive) trap state in the ZnSe:Mn particle. Chapter 7 shows the first colloidal luminescent doped nanowires. These doped luminescent ZnSe:Mn nanowires were formed by oriented attachment. The possibility to align these anisotropic structures, creates the opportunity to study the optical properties of substitutional Mn dopants and of zinc blende ZnSe exciton emission from oriented crystals.

1.6 Semiconductor Nanocrystal Applications 1.6.1 Introduction on Applications The fascinating and unique properties of colloidal nanosized semiconductor crystals attract enormous interest from various scientific and applied research fields. Several startup companies explore today the commercial production of nanocrystalline colloidal particles for both industrial and scientific products. The size tunable bright luminescence combined with a large absorption cross-section and the capability to modify the surface chemistry of QDs, make QDs the ideal chromophore for bio-labeling. A fast expanding collection of publications in the field of luminescent bio-labels shows the enormous potential of these novel materials, for both biological and medical applications. As an example, highly luminescent CdSe/ZnS core-shell QDs developed in this research, were applied by Mulder et al.57 in a novel luminescent MRI contrast agent. This application enabled the monitoring of angiogenesis in cancer tumors by both MRI and PL microscopy. The work in this thesis will however focus on the application of semiconductor NCs in lighting applications. Driven by both commercial and environmental arguments there is now a growing demand for new, highly efficient light sources. These developments will eventually lead to an almost complete replacement of traditional incandescent light bulbs and mercury filled fluorescent light sources by diode based solid state light sources. The tunable highly luminescent QDs can play an important role in these novel lighting systems as inorganic tunable chromophores. The semiconductor NCs discussed in this thesis are both applied in photoluminescent (Chapter 2 and 6) and electroluminescent applications (Chapter 2, 4, and 5) in order to create light sources that exhibit distinct optical properties, such as linear polarization of the emitted light from a

14 Chapter 1

thin film hybrid light emitting diode (LED), or from fluorescent nanocrystalline phosphors. 1.6.2 Hybrid Organic and NC-LEDs The choice to use the organic LED as a platform to study electroluminescence from NCs follows from the fact that colloidal NCs allow a similar processing behavior as organic polymers, as result of the organic capping and tunable surface properties. The flexibility of solution processing with both NCs dispersion and polymers enables the creation of nearly complete solution processed thin film LEDs. NC-LEDs can therefore be regarded as an extension to the well known family of from the organic and polymer LEDs (O-LEDs and P-LEDs). The definition of electroluminescence (EL) is: “the generation of light by electrical excitation”, and was first reported for an organic material in the 1960s with anthracene crystals.58 It was found that the process responsible for EL requires injection of electrons in the emissive material from one electrode and injection of holes at the opposite electrode.59 The oppositely charged carriers move over a certain distance in the material until recombination takes place (exciton formation), followed by exciton decay under emission of light. Organic based LEDs are constructed by the use of thin films of conjugated molecules or polymers sandwiched between a high and a low work function electrode. The polymers and molecules commonly used in these devices derive their semiconducting and emissive properties from delocalized π-orbitals. Electrodes with different work functions are needed for efficient injection of the charge carriers in the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the polymer or molecule. The organic material in these LEDs takes care of charge carrier transport and light emission.

Figure 9: Schematic energy level diagram of an organic LED under forward bias

Introduction 15

The operating principles of the organic LEDs are very different compared to the traditional inorganic LEDs. The rectification and light emission in inorganic LEDs results from the interface between the oppositely doped p and n type semiconductor, while the diode behavior in organic LEDs originates from the asymmetric electrode contacts, which give rise to a built-in potential and a small (or negligible) injection barrier for only one type of charge carrier. Under forward bias, a voltage higher than the built-in potential will cause electrons to be injected into the LUMO at the cathode and holes in the HOMO at the anode. Under the influence of the electric field inside the device (Figure 9), the electrons and holes will drift towards the opposite electrode, forming an exciton when they meet. Due to spin statistics, triplet and singlet excitons will be formed. For simple organic molecules and polymers only singlet excitons are emissive. The presence of NCs containing “heavy” metal atoms creates mixing between triplet and singlet states and is expected to increase the formation of emissive excitons in NC-LEDs compared to O-LEDs. In Chapter 2 core-shell QDs are used in QD-LEDs. The QDs in these composite devices are the emissive component, dispersed in a polyspirofluorene (PSF) matrix for charge transport. The QDs in this system were excited by both energy transfer from the PSF matrix and by carrier trapping. A sandwich like NC-LED configuration is used in Chapters 4 and 5. The diodes in Chapter 4 contained a thin layer of CdTe/CdS QDs sandwiched between an organic hole and electron transport layer, to balanced so that hole and electron transport to cause recombination in the QD layer. Aligned QRs were used in a similar LED sandwich structure (Chapter 5), in order to create a QR-LED with linearly polarized emission.

1.7 Aim of the Thesis

The overview in this chapter highlights some of the crucial steps in the recent development of a novel class of luminescent nanomaterials. The last decades have witnessed an enormous progress in both theoretical understanding and experimental work that lead to an outstanding degree of control over the particle size, shape, and monodispersity. Several successful applications of these nanosize semiconductors have been shown in bio-imaging, photovoltaics, and LEDs.

The aim in this research was to design, synthesize, characterize, and construct novel luminescent materials and electroluminescent devices. The challenges are to create these materials and devices. The main topics in this thesis are energy transfer in nanocrystalline solids, the synthesis of highly efficient NCs for luminescent applications

16 Chapter 1

with extended optical and electronic properties, and to use these novel type of luminescent materials in an actual device to create light sources with distinct optical properties.

Highly luminescent cadmium and zinc based nanoparticles were created that derive their distinct optical properties from high control over size, shape, and composition. By the combination of different materials both heterostructures and doped nanomaterials were made. The combination of size, shape, and doping effects in these nanomaterials resulted in the creation of several novel materials with a high potential for bio-, lighting, and spin related applications. Optical spectroscopy was used to study the energy transfer between different nanoparticles and the host matrix. This resulted eventually in the fabrication of both a nanophosphor with isotropic and anisotropic optical properties and the creation various electroluminescent devices.

References 1 Wertheim, G. K. Zeitschrift für Physik D Atoms, Molecules and Clusters 1988, 12, 319 2 Leonhardt, U. Nature Photonics 2007, 1 , 207 3 Fraday, M. Philos. Trans. 1857, 147, 145 4 Berry, C. R.; Phys. Rev. 1967, 153, 989 5 Berry, C. R.; Phys. Rev. 1967, 161, 848 6 Evans, B. L.; Young, P. A. Proc. R. Soc. London Ser. A. 1967 298, 74 7 Brus, L. J. Phys. Chem. 1986, 90, 2555 8 Steigerwald, M. L.; Alivisatos, A. P.; Gibson, J. M.; Harris, T. D.; Kortan, R.;

Muller, A. J.; Thayer, A. M.; Duncan, T. M.; Douglass, D. C.; Brus, L. E. J. Am. Chem. Soc. 1988, 110, 3046

9 Brennan, J. G.; Siegrist, T.; Carroll, P. J.; Stuczynski, S.M.; Brus, L. E. Steigerwald, M. L. J. Am. Chem. Soc. 1989, 111, 4141

10 Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc.1993, 115, 8706. 11 Pan, D.; Wang, Q.; Jiang, S.; Ji, X.; An, L. Adv. Mater. 2005, 17, 176 12 Deng, D. W.; Qin, Y. B.; Yang, X.; Yu, J. S.; Pan, Y. J. Cryst. Growth 2006, 296,

141 13 Gaponik, N.; Talapin, D. V.; Rogach, A. L.; Hoppe, K.; Shevchenko, E. V.;

Kornowski, A.; Eychmüller, A.; Weller, H. J. Phys. Chem. B 2002, 106, 7177 14 Talapin, D. V.; Rogach, A. L.; Kornowski, A.; Haase, M.; Weller, H. Nano Lett.

2001,1, 207 15 Dance, I. G.; Choy, A.; Scudder, M. L. J. Am. Chem. Soc. 1984, 106, 6285 16 Dingle, R. Festkörperploblem 1975, XV, 21-48

Introduction 17

17 Hu, J. T.; Li, L. S.; Yang, W. D.; Manna, L.; Wang, L. W.; Alivisatos, A. P.

Science 2001, 292, 2060 18 Chen, X.; Nazzal, A. Y.; Xiao, M.; Peng, Z. A.; Peng, X. G . J. Lumin. 2002, 97, 205 19 Mohamed, M. B.; Burda, C.; El-Sayed, M. A. Nano Lett. 2001, 1, 589 20 Manna, L.; Scher, E. C.; Alivisatos, A. P. J. Clust. Sci. 2002, 13, 521 21 Shabaev, A.; Efros, A. L. Nano Lett. 2004, 4, 1821 22 Klimov, V. I. Semiconductor and Metal Nanocrystals, Marcel Dekker Inc. 2004, 24 23 Peng, X. G. Adv. Mater. 2003, 15, 459 24 Yu, W. W.; Wang, Y. A.; Peng, X. G. Chem. Mater. 2003, 15, 4300 25 Manna, L.; Wang, L. W.; Cingolani, R.; Alivisatos, A. P. J. Phys. Chem. B 2005,

109, 6183 26 Niederberger, M.; Cölfen, H. Phys. Chem. Chem. Phys. 2006, 8, 3271 27 Pradhan, N.; Xu, H., Peng, X. Nano Lett. 2006, 6, 720 28 Cho, K. S.; Talapin D. V. Gaschler, W.; Murray, C. B. J. Am. Chem. Soc. 2005, 127, 7140 29 Korgel, B. A.; Fitzmaurice, D. Adv. Mater. 1998, 10, 661 30 Tang, Z.; Kotov, N. A.; Giersig, M. Science 2002, 297, 237 31 Pacholski, C.; Kornowski, A.; Weller H. Angew. Chem., Int. Ed. 2002, 41,1188 32 Shim, M.; Guyot-Sionnest, P. J. Chem. Phys. 1999, 111, 6955. 33 Erwin, S. C.; Zu, l.; Haftel, M. L.; Efros, A. L.; Kennedy, T. A.; Norris, D. J.

Nature 2005, 436, 91 34 Fu, H.; Zunger, A. Phys. Rev. B 1996, 56, 1469 35 Steckel, J. S.; Zimmer, J. P.; Coe-Sullivan, S.; Stott, N. E.; Bulović, V.; Bawendi, M. G. Angew. Chem. Int. Ed. 2004, 43, 2154 36 Peng, X.; Schlamp, M C.; Kadavanich, A. V.; Alivisatos, A. P. J. Am. Chem. Soc.

1997, 119, 7019 37 Hines, M. A.; Guyot-Sionnest P. J. Phys. Chem. 1996, 100, 468 38 Dabbousi, B. O.; Rodriquez-Viejo, J.; Mikulec, F. V.; Heine, J. R.; Mattoussi, H.; Ober, R.; Jensen, K. F.; Bawendi, M. G. J. Phys. Chem. B 1997, 101, 9463 39 Mekis, I.; Talapin, D. V.; Kornowski, A.; Haase, M.; Weller, H. J. Phys. Chem. B, 2003, 107, 7454 40 Reiss, P.; Bleuse, J.; Pron, A. Nano Lett. 2002, 2, 781 41 Talapin, D. V.; Mekis, I.; Götzinger, S.; Kornowski, A.; Benson, O.; Weller, H. J.

Phys. Chem. B 2004, 108, 18826

18 Chapter 1

42 Balet, L. P.; Ivanov, S. A.; Piryatinski, A.; Achermann, M.; Klimov, V. I. Nano

Lett. 2004, 4, 1385 43 Kim, S.; Fisher, B.; Eisler, H.-J.; Bawendi, M. J. Am. Chem. Soc. 2003, 125, 11466 44 Yu, K.; Zaman, B.; Romanova, S.; Wang, D.-S.; Ripmeester, J. A. Small 2005, 1, 332 45 Halpert, J. E., Porter, V. J.; Zimmer, J. P.; Bawendi, M. G. J. Am. Chem. Soc. 2006, 128, 12590 46 Milliron, D. J.; Hughes, S. M.; Cui, Y.; Manna, L.; Li, J.; Wang, L. W.; Alivisatos

A. P. Nature 2004, 430, 190 47 Talapin, D. V.; Nelson, J. H.; Shevchenko, E. V.; Aloni, S.; Sadtler, B.; Alivisatos,

A. P. Nano. Lett. 2007, 7, 2951 48 Chin, P. T. K.; de Mello Donegá, C.; van Bavel, S. S.; Meskers, S. C. J.;

Sommerdijk, N. A. J. M.; Janssen R. A. J. J. Am. Chem. Soc. 2007, 129, 14880 49 Yang, Y. A.; Wu, H.; Williams, K. R.; Cao, Y. C. Angew. Chem., Int. Ed. 2005, 44, 6712. 50 Rogach, A. L.; Kornowski, A.; Gao, M.; Eychmuller, A.; Weller, H. J. Phys. Chem.

B 1999, 103, 3065 51 Manna, L.; Milliron, D.; Meisel, A.; Scher, E. C.; Alivisatos, A. P. Nature Mater.

2003, 2, 382 52 Xie, R.; Kolb, U.; Basche, T. Small 2006, 2, 1454 53 Santangelo, S. A.; Hinds, E. A.; Vlaskon, V. A.; Archer, P. I.; Gamelin D. R. J. Am.Chem. Soc. 2006, 129, 3973 54 Raola, O. E.; Strouse G. F. Nano Lett. 2002 2, 1443 55 Norris, D. J.; Yao, N.; Charnock, F. T.; Kennedy, T. A. Nano Lett. 2001, 1, 3 56 Pradhan, N.; Peng, X. J. Am. Chem. Soc. 2007, 129, 3339 57 Mulder, W. J. M.; Koole, R.; Brandwijk, R. J.; Storm, G.; Chin, P. T. K.; Strijkers,

G. J.; de Mello Donega, C.; Nicolay, K.; Griffioen A. Nano Lett. 2006 16, 1 58 Pope, M.; Kallmann, H.; Magnante, P. J. Chem. Phys. 1963, 38, 2042 59 Burroughs, J. H.; Bradley, D. D. C.; Brown, A. R.; Marks, R. N.; Mackey, K.;

Friend, R. H.; Burn, P. L. Holmes, A. B. Nature 1990 347, 539

19

Chapter 2

Energy Transfer in Hybrid Quantum Dot LEDs

Summary Energy transfer in a host-guest system consisting of a blue-emitting poly(2,7-spirofluorene) (PSF) donor and red-emitting CdSe/ZnS core shell quantum dots (QDs) as acceptor is investigated in solid films, using time-resolved optical spectroscopy, and in electroluminescent diodes. In the QD:PSF composite films the Förster radius for energy transfer is found to be 4-6 nm. In electroluminescent devices lacking an electron transport layer, the electroluminescence (EL) spectrum of the QD:PSF polymer composite is similar to the photoluminescence (PL), giving evidence for energy transfer from PSF to the QDs. The addition of an electron transport layer between the emitting layer and the cathode results in a significant change in the EL spectrum and a considerable improved device performance, providing almost pure monochromatic emission at 630 nm with an luminance efficiency of 0.32 cd/A. The change in spectrum signifies that the electron transport layer changes the dominant pathway for QD emission from energy transfer from the polymer host to direct electron-hole recombination on the QDs.

20 Chapter 2 2.1 Introduction High quality colloidal core-shell semiconductor nanocrystals, or quantum dots (QDs), offer tunable narrow and intense photoemission as function of size in the visible range1-7 as a result of the spatial confinement of the excited charge carriers.8,9 This property can be used to make hybrid QD organic polymer light-emitting diodes (QD-LEDs) that combine the emitting properties of QDs with the flexibility in device construction of the organic and polymer materials. The use of QDs as a replacement of organic, polymer, or organometallic chromophores in LEDs has been demonstrated and is attracting increasing interest in an effort to obtain devices that combine the advantages of both systems for monochromatic visible and near infrared emission as well as for creating white light.10-41 Despite recent progress, device efficiencies of QD-LEDs still lag behind the more common organic and polymer LEDs.

Two types of QD-LED architectures can be discriminated. In the first device layout a thin QD layer is sandwiched between a hole and electron injection layer such that excitons are formed directly in the QD layer.10-28 In the second layout, the active layer consists of a blend of QDs dispersed in a polymer29-39 or small molecule matrix.40,41 The QDs in this composite material serve as emissive traps for (migrating) excitons that are generated in the polymer matrix by charge carrier recombination. The use of such hybrid system where the QDs are embedded in a polymer matrix generally gives low luminance efficiency (~0.05 cd/A) for monochromatic devices but was recently reported to be 2.2 cd/A for white-light-emitting devices.40 In these QD-LEDs, the QD electroluminescence (EL) originates either from recombination of injected charges in the host followed by Förster energy transfer33,35,42-44 to the QD, or by direct trapping and recombination of injected charge carriers on the QDs. In photoluminescence (PL), on the other hand, no (or few) free charge carriers are created in the host after photoexcitation and QD emission mainly stems from Förster energy transfer from the host, or from direct excitation of the QD. An in-depth study on energy transfer and carrier trapping differences in PL and EL in QD/polymer composite LEDs can contribute to the improvement of hybrid QD LEDs.

In this study we use a conjugated blue-emitting (450 nm) poly(2,7-spirofluorene) (PSF) that possesses a fluorescence quantum yield of 40% as the host polymer matrix and energy donor together with red-emitting (630 nm) CdSe/ZnS core shell QDs as energy acceptor. We show that the photoluminescence of the PSF polymer and the CdSe/ZnS core-shell QDs in mixed films is governed by energy transfer from PSF to QDs. The mechanism can be described by Förster theory assuming a Förster radius of 4-6 nm. The results obtained from photoexcitation are compared with electroluminescence studies of the same layers. In these QD-LEDs energy transfer plays an important role when charge recombination is dominant in the polymer but by introducing an electron transport layer,

Energy Transfer in Hybrid Quantum Dot LEDs 21

the QD emission can be significantly enhanced as a consequence of direct electron-hole recombination, leading to a red-light-emitting device with increased luminance efficiency.

2.2 Experimental Materials and sample preparation. The poly(2,7-spirofluorene) (PSF) was obtained from Covion Organic Semiconductors GmbH.45,46 CdSe/ZnS QDs were prepared according to literature procedures.2 Poly(3,4-ethylenedioxythiophene):poly-(styrenesulfonate) (PEDOT:PSS), high resistance PEDOT 5411 Baytron was obtained from Bayer AG. TPBI (1,3,5-tris(N-phenylbenzimidazol-2-yl)benzene) was obtained from Sensient Imaging Technologies Gmbh. All solvents were of analytical quality. The QDs were purified two times by dissolving a solid powder of singly purified CdSe/ZnS QDs in a certain amount of chloroform to obtain a 1% (w/V) dispersion and precipitating with an equal amount of methanol. The QDs were collected by centrifugation, and dissolved in chloroform. Mixtures of the QD:PSF solutions in chloroform were deposited by spin coating, using a BLE Delta 20 BM spin coater. For photoluminescence measurements the emissive layer is spin coated on clean quartz substrates. Optical spectroscopy. Steady state photoluminescence spectra were recorded using a Perkin–Elmer LS 50B spectrometer using 4.6 eV as the excitation energy. UV-vis spectra were recorded using a Perkin–Elmer Lambda 900 spectrophotometer. Time-resolved fluorescence was measured using a streak camera set-up (Chromex 250is Polychromator 40 groves/mm grating, Hamamatsu 5677 Slow Speed Sweep Unit) in the dump mode with a temporal resolution of about 2 ps in the 2 ns detection window. The resolution in the detection window of 12 ns was 0.12 ns. The excitation was carried out at 380 nm (Spectra Physics Millenia Xs pump laser, Spectra Physics Tsunami mode-locked Ti:sapphire laser, Spectra Physics 3980 frequency doubler and pulse selector). The streak camera spectra were corrected for the spectral response of the incoupling lenses, the polychromator, the streak tube, and the shading effects due to the deflection plate. Device preparation and characterization. The QD light-emitting diodes (QD-LEDs) were fabricated under clean room conditions, using patterned ITO/glass substrates with a 120 nm thick transparent ITO layer as the bottom electrode. The ITO/glass substrates are treated for 15 min with UV/ozone (UVP PR-100) before processing. A ~100 nm PEDOT:PSS layer was deposited by spin coating and annealed at 180 °C for 2 min. Subsequently the emissive QD:PSF mixture was deposited from chloroform solution by spin coating. The TPBI layer (40 nm) and Ba (5 nm)/Al (100 nm) metal cathode were deposited by vacuum evaporation. The device area was 0.09 cm2. The QD-LEDs were characterized using a low-noise single channel DC power source, using a voltage/current

22 Chapter 2 source meter (Keithley 2400, Keithley Instruments). Light from the LED was measured using a photodiode and read out by an electrometer/high-resistance meter (Keithley 2400). The photodiode was calibrated with a luminance meter (Minolta LS-110). The electroluminescence spectra were recorded using a fiber-coupled spectrograph/CCD camera combination (Ocean Optics S2000). The emission was corrected for the wavelength dependence of the spectrometer.

2.3 Results and Discussion 2.3.1 Energy Transfer The absorption and photoluminescence (PL) spectra of CdSe/ZnS QDs in chloroform solution is shown in Figure 1 and compared to the fluorescence spectrum of PSF. Figure 1 reveals that the QD absorption spectrum has a significant overlap with the fluorescence of PSF. This overlap is a requirement to enable efficient energy transfer from PSF to the QDs when they are mixed,47,48 and the spectral separation between PSF and QD PL emission allows detecting both processes independently.

400 450 500 550 600 650 7000

100

200

300

400

500

0.0

0.1

0.2

0.3

0.4

0.5

0.00

0.05

0.10

0.15

0.20

Pho

tolu

min

esce

nce

(cou

nts)

Wavelength (nm)

QD Abs

QD PLPSF PL

n

R'R

Absorbance

Figure 1. Absorption (open squares) and PL (solid triangles) spectra of CdSe/ZnS QDs compared to the PL spectrum of PSF (solid squares). All spectra were recorded for chloroform solutions at room temperature. The inset shows the molecular structure of PSF.

Energy Transfer in Hybrid Quantum Dot LEDs 23

The efficiency of energy transfer from the PSF donor to the CdSe/ZnS QD acceptor can be expressed by the Förster radius (R0) at which half of the excited donor molecules decay by energy transfer and the other half by intrinsic radiative and non radiative pathways. When energy transfer takes the form of interacting transition dipole moments on donor and acceptor, the Förster distance can be estimated from the spectral overlap J (in nm4/Mcm) of the photoluminescence (FD(λ)) of the donor and the absorption (εA(λ)) of the acceptor, via: 47,48

6/1F

420 ])D([211.0 JnR ηκ −= (in Å) (1)

where κ2 accounts for the relative orientation of the two transition dipole moments and is assumed to be equal to 2/3 for random orientation of the dipole moments.49 ηF(D) is the luminescence quantum yield of the donor in the absence of acceptor, and n is the refractive index of the solvent. Using the spectral overlap obtained from the spectra shown in Figure 1, the Förster radius for PSF and CdSe/ZnS was determined to be ~6.2 nm, in agreement with the values 5.4-5.8 nm42 and 6.7-7.0 nm44 that were recently reported for similar combinations of CdSe/ZnS QDs and a wide band gap semiconducting polymer. Hence, in this range energy transfer from PSF to CdSe/ZnS QDs is rather efficient. To investigate the energy transfer in films, the QDs were mixed with PSF in different mass ratios and deposited by spin coating from chloroform on quartz substrates. Figure 2a shows that the PL intensity of PSF in these mixed QD:PSF films decreases significantly with increasing QD concentration. At the same time, the PL intensity of the QDs increases, consistent with the expected energy transfer, but possibly also because of direct excitation. The PL excitation spectrum recorded at the maximum of the QD emission (630 nm) for the 70 wt.% QD:PSF blend, however, shows the characteristic features of the absorption of PSF (Figure 2b) and confirms that when exciting at ~400 nm the QD emission results mainly from energy transfer (ET) from PSF to the QDs. The low intensity tail in the PL excitation above ~450 nm (Figure 2b) is due to absorption by the QDs

24 Chapter 2

400 450 500 550 600 6500

200

400

600

800

1000 0 20 40 50 70 90

Pho

tolu

min

esce

nce

(a.u

.)

Wavelength (nm)

increasing QD concentration

(a)

350 400 450 500 5500

100

200

300

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Pho

tolu

min

esce

nce

(a.u

.)

Wavelength (nm)

pure QD

70%

absorption PSF

(b)

Absorbance

Figure 2. (a) PL spectra of QD:PSF composite films for different wt.% of QDs (see inset) in the film. The PL intensity has been corrected for the absorbance at the excitation wavelength (270 nm). (b) PL excitation spectrum of a 70 wt.% QD:PSF film recorded at 630 nm (solid triangles) together with the absorption spectrum of PSF (open squares) and the PL excitation spectrum of a pure QD film (open circles). For Förster energy transfer from a donor (PSF) to an acceptor (QD) that is randomly but rigidly distributed in three dimensions, the fluorescence intensity of the donor in donor-acceptor mixture (IDA) can be described by:48

)](1[12

D

DA γγπ γ erfeI

I−=− (2)

Energy Transfer in Hybrid Quantum Dot LEDs 25

Where γ is given by:

30a 3

42

RC ππγ = (3)

and ID is the donor emission intensitie in the absence of the acceptor and Ca the concentration of QD acceptors (mol/nm3). To estimate R0, different CdSe QDs batches were studied that had similar size and composition (2 or 3 monolayers of ZnS) and optical properties (λem ≈ 630 nm). The relative quenching (1-(IDA/IA)) of the donor (PSF) fluorescence as function of Ca is plotted Figure 3 and compared to the calculated curves for different values for R0. The experiments shown in Figure 3 represent two different batches of QDs, each incorporated in two QD/polymer films, resulting in four sets represented by different markers. As can be seen there is a considerable spread in the experimental data, due to inhomogeneous film formation, but the general trends shown in Figure 3 are consistent with eq 2. when R0 is in the range of 4-6 nm, in fair agreement with the 6.2 nm estimated from spectral overlap between donor emission and acceptor absorption.

10-4 10-3

0.0

0.2

0.4

0.6

0.8

1.0

R0 (nm) 6.0 5.0 3.8

1-I D

A / I D

Ca (mol/nm3)

Figure 3. Relative quenching (1-(IDA/IA)) of the donor (PSF) fluorescence as function of the quantum dot concentration (Ca) in the film. The lines represent eq. 2 for different Förster distances R0. The experimental data are obtained for two different batches of QDs, each measured in two sets of experiments.

26 Chapter 2 Figure 4a shows the time-resolved photoluminescence intensity recorded at 460 nm of pristine PSF and of mixtures of QDs in PSF (20 and 50 wt.%). The fluorescence of PSF can be described by a bi-exponential decay with lifetimes τ1 = 95±1 ps and τ2 = 580±4 ps, with relative weights of about 2:1. As expected for energy transfer, the addition of QDs results in a decrease in emission lifetime of PSF (τ1 = 83±1 ps and τ2 = 452±3 ps with relative weight of 3:1 for 20 wt.% QDs, and τ1 = 73±1 ps and τ2 = 440±3 with relative weight of 4:1 for 50 wt.% QDs). Figure 4b shows the QD time-resolved luminescence intensity monitored at 630 nm of the pure QDs and two QD:PSF blends. For the pure QDs the rise is mono-exponential with a time constant of ~3.7 ps which is half of the FWHM of the machine response (8 ps). For the mixed films we find a bi-exponential growth of the QD emission. The rise of the QD emission in QD:PSF blends clearly shows a contribution at longer timescales which we attribute to energy transfer from PSF to QD. For the 20 wt.% blend, the QD emission rises with τ1 = 6±1 ps and τ2 = 39±3 ps, while for the 50 wt.% blend the characteristic times are τ1 = 14±1 ps and τ2 = 232±15 ps. In both cases we attribute the short time to result mainly from direct QD excitation, while the long time is a typical signature of the energy transfer.

Energy Transfer in Hybrid Quantum Dot LEDs 27

0 500 1000 1500 2000102

103

0 wt.% 20 wt.% 50 wt.%

Inte

nsity

(cou

nts)

Time (ps)

(a)

0 200 400 600 800

200

400

600

800

1000

1200

1400

1600

0

2

4

6

8

10

12

14

Inte

nsity

(a.u

.)

Time (ps)

50 wt.%

20 wt.%

100 wt.%

(b)

Figure 4. Time-resolved photoluminescence. (a) PSF emission at 460 nm of pure PSF (solid squares, 0 wt.% QD) and of QD:PSF blends with 20 and 50 wt.% QDs. (b) QD emission at 630 nm of pure QDs (100 wt.%) and QD:PSF blends 20 and 50 wt.% QDs. The red lines represent fits of a bi-exponential rise to the experimental data. The blue line represents the machine response of the excitation pulse. 2.3.2 Electroluminescence Electroluminescence (EL) was measured for composite QD:PSF films sandwiched between an indium tin oxide (ITO)/poly(3,4-ethylenedioxythiophene):poly-(styrenesulfonate) (PEDOT:PSS) anode and a Ba/Al cathode. The device architecture

28 Chapter 2 (Figure 5a) contains an optional 1,3,5-tris(N-phenylbenzimidazol-2-yl)benzene (TPBI) electron transport layer (ETL). The energy diagram of these QD-LEDs is shown in Figure 5b and reveals that in the active layer holes will be confined to PSF while electrons may become trapped on the CdSe core.

ITO PEDOT PSF ZnS CdSe ZnS PSF TPBI Ba Al-8

-7

-6

-5

-4

-3

-2

-1

0

(a)

(b)

2.7

4.4

3.4 3.4

2.02.0

4.3

2.5

6.76.57.0 7.0

5.25.24.75.0E

nerg

y (e

V)

Figure 5. (a) Schematic of the QD-LED device structure. The TPBI layer was not used in all devices (see text). (b) Energy levels of the various materials with respect to vacuum. In first approximation, the EL spectra of the QD:PSF composite film QD-LED devices without TPBI layer (Figure 6a) are similar to the corresponding PL spectra (Figure 2). The highest QD EL intensity is found for the layer containing 60 wt.% QDs.

Ba/AlTPBI

QD:PSFPEDOT:PSS

ITOGlass

VBa/AlTPBI

QD:PSFPEDOT:PSS

ITOGlass

V

Energy Transfer in Hybrid Quantum Dot LEDs 29

400 500 600 7000.0

0.1

0.2

0.3

0.4

0.5(a)

20 40 60 80 90

Ele

ctro

lum

ines

cenc

e (a

.u.)

Wavelength (nm)

400 500 600 7000.00

0.05

0.10

0.15

0.20

0

5

10

15

20

25

30(b)

EL TPBI no TPBI

Elec

trolu

min

esce

nce

(a.u

.)

Wavelength (nm)

80 wt.% QD

PL

Photolum

inescence (a.u.)

0 2 4 6 8 10 12 14

0

20

40

60

80

100

120

140

6 8 10 12 14 160.01

0.1

1

10

(c) TPBI 80% TPBI 70% no TPBI 80%

Cur

rent

den

sity

(A/m

2 )

Bias (V)

Bias (V)

Lum

ines

cenc

e (c

d/m

2 )

Figure 6. (a) EL spectra of ITO/PEDOT:PSS/QD:PSF/Ba/Al QD-LEDs for different concentrations of QDs (in wt.%) measured at J = 55 A/m2. (b) EL spectrum of an ITO/PEDOT:PSS/QD(80 wt.%):PSF/TPBI/Ba/Al QD-LED (open triangles). The closed triangles show the corresponding EL spectrum without the TPBI ETL. The solid line represents the PL spectrum of the same film. (c) Current density and luminance of ITO/PEDOT:PSS/QD:PSF/Ba/Al LEDs versus the bias voltage without (solid symbols) and with (open symbols) a TPBI layer.

30 Chapter 2 When the QD emission intensity in the blends is compared to that of PSF for the EL and PL experiments (Figure 7), the increase of relative intensity with QD concentration is similar within experimental error. This similarity suggests that energy transfer from PSF to the QDs is responsible for the EL of the QDs and that direct electrical excitation (e-h recombination) on the QDs is not predominant in these devices.

0 20 40 60 800.0

0.5

1.0

1.5

2.0

2.5

3.0

I(630

) / I(

455)

wt.% QD

PL EL

Figure 7. Relative intensities of QD (630 nm) and PSF (455 nm) emission intensity from photoluminescence (open markers) and electroluminescence (closed markers) versus the concentration of QDs. At high QD wt.% the PSF emission becomes very small and the error in the ratio increases. The performance of LEDs strongly depends on the balance of hole and electron currents. When the mobilities of holes and electrons differ significantly, an imbalance of charge carriers in the emitting layer will result. The excess of one type of charge carriers will lower the device performance because charge carriers may pass the active layer without recombination. Confinement of charge carriers to the emitting layer can be achieved by introducing electron or hole blocking layers. To confine holes in the light-emitting QD:PSF layer we introduced a 40 nm thick thermally evaporated TPBI electron transport / hole blocking layer (ETL/HBL) between the QD:PSF layer and the Ba/Al cathode (Figure 5a). An additional advantage of an ETL/HBL is that it minimizes exciton quenching at the Ba/Al cathode. Excitons close to the metal electrode often decay non-radiatively. Figure 6b shows the EL spectrum obtained for QD-LEDs with 80 wt.% QDs in PSF (open markers). The 40 nm TPBI layer results in an increase in QD emission intensity by more than one order of magnitude compared to the device without TPBI

Energy Transfer in Hybrid Quantum Dot LEDs 31

(solid markers), while the polymer emission exhibits a threefold increase in EL intensity. The larger increase in the QD emission compared to the PSF emission reveals that the TPBI layer causes charge recombination on the QDs to become the dominant pathway for exciting the QDs. The presence of polymer emission in the device with a TPBI layer is indicative of exciton formation in PSF, which is even slightly increased by the TPBI layer as result of improved charge and exciton confinement. This shows that the QDs are not solely excited by charge trapping on the QDs but that energy transfer from PSF still occurs. The current density and luminance of the devices with and without the TPBI layer are shown in Figure 6c. As can be seen, the current density of the QD-LEDs exhibits some sudden changes that are reminiscent of resistive switching phenomena as observed in CdSe QD based organic memories.50,51 The 40 nm thick TPBI causes an increase in onset voltage for the current but not for light output. As a consequence, the QD-LED (80 wt.% QD in PSF) without TPBI layer has a maximum luminance efficiency of only 0.015 cd/A, which is increased to 0.16 cd/A when using TPBI. The best device in terms of luminance efficiency (0.32 cd/A, Figure 6c) was obtained for a QD-LED with a slightly lower concentration of QDs (70 wt.% in PSF) that included a TPBI electron transport layer to enhance the QD emission compared to the PSF emission, similar to the 80 wt.% blend shown in Figure 6c.

2.4 Conclusion Photoluminescence spectroscopy reveals that energy transfer in blends of core-shell CdSe/ZnS QDs and PSF as a conjugated polymer can be described with an average Förster radius between 4 and 6 nm, in agreement with the estimate (6.2 nm) determined from the spectral overlap between donor emission and acceptor absorption. Energy transfer from PSF to the QDs was also evidenced from the PL excitation spectra and is reflected in the luminescence intensity dynamics where the QD emission continues to increase after the excitation pulse. The electroluminescence spectra of QD:PSF composite layers are similar to the photoluminescence spectra for devices that do not use a TPBI electron transport layer. The similarity suggests that under these conditions the QD emission arises mainly from Förster energy transfer from PSF and that direct electrical excitation (e-h recombination) on the QDs is not predominant. By using a TPBI electron transport layer the electroluminescence spectrum was tunable to a more pure monochromatic QD emission. The considerably enhanced QD emission resulted in devices with luminance efficiency of 0.16 and 0.32 cd/A, for 80 and 70 wt.% QDs in PSF, respectively. The improved device performance together with the significantly increased QD emission suggests that TPBI enhances the emission that originates predominantly

32 Chapter 2 from direct electron-hole recombination in the QDs, by improved chare carrier trapping and by exciton confinement in the emissive layer. The results show that electroluminescence in QD composite LEDs does not mainly depend on energy transfer, but also on direct carrier recombination.35,40 In electroluminescence, carrier trapping becomes the main pathway for excitation in QDs when the charge carriers are effectively confined to the emissive layer.

References 1 Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993, 115, 8706 2 Talapin, D. V.; Rogach, A. L.; Kornowski, A.; Haase, M.; Weller, H. Nano Lett.

2001, 1, 207 3 de Mello Donegá, C.; Hickey, S. G.; Wuister, S. F.; Vanmaekelbergh, D.;

Meijerink, A. J. Phys. Chem. B 2003, 107, 489 4 Peng Z. A.; Peng, X. J. Am. Chem. Soc. 2001, 123, 1389 5 Dabbousi, B. O.; Rodriguez-Viejo, J.; Mikulec, F. V.; Heine, J. R.; Mattoussi, H.;

Ober, R.; Jensen, K. F.; Bawendi, M. G. J. Phys. Chem. B 1997, 101, 9463 6 Mekis, I.; Talapin, D. V.; Kornowski, A.; Haase, M.; Weller, H. J. Phys. Chem. B

2003, 107, 7454 7 Reiss, P.; Bleuse, J.; Pron, A. Nano Lett. 2002, 2, 781 8 Alivisatos, A. P. J. Phys. Chem. 1996, 100, 13226 9 Efros A. L.; Rosen, M. Annu. Rev. Mater. Sci. 2000, 30, 475 10 Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994, 370, 354 11 Schlamp, M. C.; Peng, X.; Alivisatos, A. P. J. Appl. Phys. 1997, 82, 5837 12 Mattoussi, H.; Radzilowski, L. H.; Dabbousi, B. O.; Thomas, E. L.; Bawendi, M.

G.; Rubner, M. F. J. Appl. Phys. 1998, 69, 377 13 Mattoussi, H.; Radzilowski, L. H.; Dabbousi, B. O.; Fogg, D. E.; Schrock, R. R.;

Thomas, E. L.; Rubner, M. F.; Bawendi, M. G. J. Appl. Phys. 1999, 86, 4390 14 Coe, S.; Woo, W. K.; Bawendi, M. G.; Bulović, V. Nature 2002, 420, 800 15 Coe-Sullivan, S.; Woo, W. K.; Steckel, J. S.; Bawendi, M. G.; Bulović, V. Org.

Electron. 2003, 4, 123 16 Hikmet, R. A. M.; Talapin, D. V.; Weller, H. J. Appl. Phys. 2003, 93, 3509 17 Steckel, J. S.; Coe-Sullivan, S.; Bulović, V.; Bawendi, M. G. Adv. Mater. 2003, 15,

1862 18 Steckel, J. S.; Zimmer, J. P.; Coe-Sullivan, S.; Stott, N. E.; Bulović, V.; Bawendi,

M. G. Angew. Chem. Int. Ed. 2004, 43, 2154 19 Chaudhary, S.; Ozkan, M.; Chan, W. C. W. Appl. Phys. Lett. 2004, 84, 2925

Energy Transfer in Hybrid Quantum Dot LEDs 33

20 Zhao, J.; Zhang, J.; Jiang, C.; Bohnenberger, J.; Basché, T.; Mews, A. J. Appl.

Phys. 2004, 96, 3206 21 Hikmet, R. A. M.; Chin, P. T. K.; Talapin, D. V.; Weller, H. Adv. Mater. 2005, 17,

1436 22 O’Conner, É.; O’Riordan, A.; Doyle, R.; Moynihan, S.; Cuddihy, A.; Redmond, G.

Appl. Phys. Lett. 2005, 86, 201114 23 Coe-Sullivan, S.; Steckel, J. S.; Woo, W. K.; Bawendi, M. G.; Bulović, V. Adv.

Funct. Mater. 2005, 15, 1117 24 Zhao, J.; Bardecker, J. A.; Munro, A. M.; Liu, M. S.; Niu, Y.; Ding, I.-K.; Luo, J.;

Chen, B.; Jen, A. K.-Y.; Ginger, D. S. Nano Lett. 2006, 6, 463 25 Steckel, J. S.; Snee, P.; Coe-Sullivan, S.; Zimmer, J. P.; Halpert, J. E.; Anikeeva,

P.; Kim, L. A.; Bulović, V.; Bawendi, M. G. Angew. Chem. Int. Ed. 2006, 45, 5796 26 Niu, Y.-H.; Munro, A. M.; Cheng, Y.-J.; Tian, Y.; Liu, M. S.; Zhao, J.; Bardecker,

J. A.; Jen-La Plante, I.; Ginger, D. S.; Jen, A. K.-Y. Adv. Mater. 2007, 19, 3371 27 Anikeeva, P. O.; Halpert, J. E.; Bawendi, M. G.; Bulović, V. Nano Lett. 2007, 7,

2196 28 Sun, Q.; Wang, Y. A.; Li, L. S.; Wang, D.; Zhu, T.; Xu, J.; Yang, C.; Li, Y. Nature

Photonics 2007, 1, 717 29 Dabbousi, B. O.; Bawendi, M. G.; Onitsuka, O.; Rubner, M. F. Appl. Phys. Lett.

1995, 66, 1316 30 Gao, M.; Richter, B.; Kirstein, S. Adv. Mater. 1997, 9, 802 31 Gao, M.; Lesser, C.; Kirstein, S.; Möhwald, H.; Rogach, A. L.; Weller, H. J. Appl.

Phys. 2000, 87, 2297 32 Tessler, N.; Medvedev, V.; Kazes, M.; Kan, S.; Banin, U. Science 2002, 295, 1506 33 Bakueva, L.; Musikhin, S.; Hines, M. A.; Chang, T. W. F.; Tzolov, M.; Scholes, G.

D.; Sargent, E. H. Appl. Phys. Lett. 2003, 82, 2895 34 Park, J. H.; Kim, J. Y.; Chin, B. D.; Kim, Y. C.; Kim, J. K.; Park, O. O.

Nanotechnology 2004, 15, 1217 35 Li, Y.; Rizzo, A.; Mazzeo, M.; Carbone, L.; Manna, L.; Cingolani, R.; Gigli, G. J.

Appl. Phys. 2005, 97, 113501 36 Tang, A. W.; Teng, F.; Xiong, S.; Gao, Y. H.; Liang, C. J.; Hou, Y. B. J.

Photochem. Photobiol. A. 2007, 92, 1 37 Xuan, Y.; Zhao, N.; Pan, D.; Ji, X.; Wang, Z.; Ma, D. Semicond. Sci. Technol.

2007, 22, 1021 38 Bertoni, C.; Gallardo, D.; Dunn, S.; Gaponik, N.; Eychmüller, A. Appl. Phys. Lett.

2007, 90, 034107 39 Xuan, Y.; Pan, D.; Zhao, N.; Ji, X.; Ma, D. Nanotechnology 2006, 17, 4966

34 Chapter 2 40 Li, Y.; Rizzo, A.; Cingolani, R.; Gigli, G. Adv. Mater. 2006, 18, 2545 41 Ahn, J. H.; Bertoni, C.; Dunn, S.; Wang, C.; Talapin, D. V.; Gaponik, N.;

Eychmüller, A.; Hua, Y.; Bryce, M. R.; Petty, M. C. Nanotechnology 2007, 18, 335202

42 Javier, A.; Yun, C. S.; Strouse, G. F. Mat. Res. Soc. Symp. Proc. 2003, 776, Q2.1.1 43 Anni, M.; Manna, L.; Cingolani, R.; Valerini, D.; Cretí, A.; Lomascolo, M. Appl.

Phys. Lett. 2004, 85, 4169 44 Kaufmann, S.; Stöferle, T.; Moll, N.; Mahrt, R. F.; Scherf, U.; Tsami, A.; Talapin,

D. V.; Murray, C. B. Appl. Phys. Lett. 2007, 90, 071108 45 van Dijken, A.; Perro, A.; Meulenkamp, E. A.; Brunner, K. Org. Electron. 2003, 4, 131 46 Becker, H.; Büsing, A.; Falcou, A.; Heun, S.; Kluge, E.; Parham, A.; Stöβel, P.; Spreitzer, H.; Treacher, K.; Vestweber, H. Proc. SPIE 2001, 4464, 49 47 Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd Ed., Kluwer Academic, 1999 48 Valeur, B. Molecular Fluorescence – An Introduction: Principles Applications,

VCH, Weinheim, 2000 49 Scholes, G. D.; Andrews, D. L. Phys Rev. B. 2005, 72, 125331 50 Das, B. C.; Batabyal, S. K.; Pal, A. J. Adv. Mater. 2007, 19, 4172 51 Sahu, S.; Majee, S. K.; Pal, A. J. Appl. Phys. Lett. 2007, 91, 143108

35

Chapter 3

Highly Luminescent CdTe/CdSe Colloidal Heteronanocrystals with Temperature

Dependent Emission Color

Summary In this work we present the preparation of highly luminescent anisotropic CdTe/CdSe colloidal heteronanocrystals. The reaction conditions used (low temperature, slow precursor addition, and surfactant composition) resulted in a tunable shape from prolate to branched CdTe/CdSe nanocrystals. Upon CdSe shell growth the hetero-nanocrystals show a gradual evolution from type-I to type-II optical behavior. These heteronanocrystals show a remarkably high photoluminescence quantum yield (up to 82%) and negligible thermally induced quenching up to temperatures as high as 373 K.

36 Chapter 3

3.1 Introduction The remarkable size and shape dependent properties of semiconductor nanostructures1 have attracted increasing interest over the past decades and, as a result, an excellent degree of control over the composition, size, shape and surface of semiconductor nanocrystals has been achieved in recent years by using colloidal chemistry routes.2

Semiconductor heteronanostructures, such as core/shell quantum dots (QDs), can show different behavior regarding charge carrier localization after photoexcitation (type-I or type-II), depending on the band offset alignments between the core and the shell materials.3 In type-I QDs both carriers are primarily confined in the same part of the heterostructure, while in type-II QDs the band offsets are such that electrons and holes are spatially separated, leading to the formation of an indirect exciton. The radiative recombination of the indirect exciton will result in emission at lower energies than those of both the core and the shell optical gaps, thus allowing access to wavelengths that would otherwise not be available with a single material or type-I QDs. Since the position of the energy levels of QDs is strongly size-dependent,1,2 the relative energy offsets can be tuned by controlling the core diameter and the shell thickness. This offers the possibility of directly controlling the degree of charge carrier localization from type-I to type-II,4 and consequently the emission wavelength, exciton radiative lifetimes and the exciton confinement energies, with important consequences for a number of potential applications such as photovoltaic devices, photocatalysts, lasers, and LED’s. The potential advantages of colloidal type-II QDs have triggered an increasing interest in the last few years,4-17 and several semiconductor combinations have been investigated as prospective type-II QDs, based on the band offsets of the bulk materials (viz., ZnSe-CdSe,4-5 CdTe-CdSe,6-12 ZnTe-CdSe6,13-15). These pioneering studies4-17 have provided a wealth of new insights into both the chemistry and the physics of colloidal type-II QDs, but there are still several challenging issues to be addressed. For example, the photoluminescence quantum yields (QYs) reported for colloidal type-II QDs are typically very low (0-10 % 6,8,10-16). High QYs (30-40%) have been reported for thin shell heteronanocrystals of both CdTe/CdSe and ZnTe/CdSe.15,16 However, these thin shell heteronanocrystals still present a type-I character, as they do not show the featureless subband-gap absorption tail indicative of the spatially indirect transitions expected for type-II QDs.6,10 With increasing shell thickness the long absorption tail appears but the transition from type-I to type-II behavior is accompanied by a strong decrease in QY to below 10%. These low QYs have been seen as an intrinsic limitation of type-II QDs,6,12 since the slower radiative recombination of indirect excitons would facilitate the dominance of nonradiative recombination at defects. However, preparation methodologies that minimize both surface and interfacial defects may substantially improve the

Luminescent CdTe/CdSe Colloidal Heteronanocrystals 37

photoluminescence QYs of type-II QDs beyond its low current values. Accordingly, a recent report describes highly luminescent CdS/ZnSe core/shell type-II nanocrystals with emission QYs up to 50% that benefit from an intermediate ZnCdSe alloy layer with a graded composition.17 Here we describe the synthesis and properties of highly luminescent CdTe/CdSe core/shell nanocrystals (NCs). The synthesis method developed in this work yields highly efficient (PL QY up to 80% at 300 K) anisotropic CdTe/CdSe core/shell NCs, with shapes tunable from prolate type-I QDs (aspect ratio: 1.5-2) to branched heteronanocrystals that exhibit type-II character. The high quality of these CdTe/CdSe heteronanocrystals is also attested by the negligible contribution of thermally induced luminescence quenching processes in the 293-383 K temperature range, allowing the temperature dependence of the emission color to be observed. The results reported here clearly demonstrate that type-II heteronanocrystals can have high photoluminescence QYs, despite the longer radiative lifetimes of indirect excitons. 3.2 Experimental Chemicals. Trioctylphosphine (TOP, 90%), trioctylphosphine oxide (TOPO, 99%), dodecylamine (DDA, 98%), hexadecylamine (HDA, 90%), and anhydrous solvents (toluene, methanol, chloroform) were all purchased from Aldrich. Selenium powder (99.99%, 200 mesh) and anhydrous cadmium acetate (CdAc, 99.99+%) were purchased from Chempur. Tellurium powder (99.999%, <250 μm) and dimethylcadmium (99.99%) were purchased from Heraeus and ARC Technologies, respectively. All chemicals were used as received, with the exception of TOPO, HDA and DDA, which were dried and degassed before use by heating under vacuum, DDA at 373 K for 6 h, TOPO and HDA at 393 K for 1.5 h. Synthesis and Experimental Details. High-quality green-emitting CdTe QDs (2.6-2.8 nm average diameter, 5-10% size dispersion) were prepared in DDA/TOP following a previously reported method18 and subsequently purified by filtration (to remove unreacted Te), precipitation with anhydrous methanol, and drying under vacuum. The purified CdTe QDs (16 mg) were dissolved in toluene (1 mL) and redispersed in a TOPO (67 wt%) – HDA (33 wt%) mixture at 393 K. Previous work has shown that this surfactant mixture allows well-controlled growth and yields high-quality CdSe QDs.19 The CdSe shell growth is carried out at 423 K by dropwise (0.05 mL/min) and alternate addition of Cd (0.86 mmol CdAc in 3 mL TOP) and Se (1.26 mmol Se in 3 mL TOP) precursors, followed by a 30 min annealing step at 403 K. Aliquots of the growth solution (Table 1) were taken at regular time intervals during the shell growth and quickly cooled into room temperature decalin. No post-preparative procedure was employed after sampling, except

38 Chapter 3

for the samples to be investigated by transmission electron microscopy, which were purified by precipitation with anhydrous methanol. Optical Spectroscopy. All optical measurements were carried out on samples with a low optical density (≤0.1 at the excitation wavelength). Absorption and PL spectra were collected with a Perkin–Elmer Lambda 900 spectrophotometer and an Edinburgh Instruments FS920 spectrophotometer, respectively. PL quantum yields were determined using fluorescein in 0.1 M NaOH20 and rhodamine 101 in 0.01% HCl ethanolic solution21 as references, following the method reported in Ref. 19. Lifetime measurements were carried out using time-correlated single photon counting (TCSPC). The nanocrystals dispersions were excited at 400 nm using a picosecond diode laser (PicoQuant PDL 800B) operated at 0.5-2.5 MHz. The detector unit consisted of a Peltier-cooled Hamamatsu microchannel plate photomultiplier (R3809U-50). Transmission electron microscopy. TEM and High-resolution TEM was performed on a TECNAI G2 20 transmission electron microscope (FEI Co., The Netherlands) operated at 200 kV or on a Titan Krios transmission electron microscope (FEI Co., The Netherlands) operated at 300 kV. The samples for TEM were prepared by dipping a carbon coated copper (400-mesh) TEM grid into a chloroform solution of nanocrystals. The excess liquid was removed by blotting using filter paper. X-ray diffraction. XRD data of the CdTe/CdSe sample was collected with a Rigaku D/Max B diffractometer (graphite monochromator, scintillation counter) using Fe Kα radiation (1.937 Å) operating at 40 kV, 30 mA. The data were collected from drop cast NCs on glass samples using a step scan mode with a step size of 0.02º and a counting time of 3 s per step in the range 2θ between 20° to 80º. XRD data from the CdTe core sample was collected with a Philips PW1729 diffractometer using Cu Kα radiation (1.5406 Å) operating at 40 kV, 20 mA. The data were collected from drop cast NCs on glass samples using a step scan mode with a step size of 0.04º and a counting time of 1 s per step in the range 2θ between 15° to 60º. For convenience the CdTe/CdSe diffraction pattern was recalculated for a Cu Kα radiation (1.54056 Å), following the “Bragg Law” nλ=2d sinθWhere n is an integer, λ is the wavelength, d is the interplanar spacing, and θ is the diffraction angle.

Luminescent CdTe/CdSe Colloidal Heteronanocrystals 39

Table 1. CdTe/CdSe Samples Isolated During Shell Growth. Reaction Time (t), Photoluminescence Maximum (λmax) and Quantum Yield (QY). CdSe Thickness along the Particle Length and Diameter Directions.

CdSe shell thickness (nm) Sample t (min) λmax (nm) QY (%) Length direction Diameter direction Core 535 7 A 5 539 34 0.2 a B 15 585 46 0.66 0.1 C 30 630 82 0.92 0.2

D 45 711 52 1.6b

E 90 789 45 6-10c 3c a Tips. b Diagonal direction assuming that the CdTe core occupies the center of the pyramids. c Branch length and diameter.

3.3 Results and Discussion 3.3.1 CdTe/CdSe Heteronanocrystal Growth The alternate addition of low reactivity precursors (such as cadmium oleate or cadmium acetate) prevents homogeneous nucleation, as has been demonstrated by several recent works employing the SILAR (Successive Ion Layer Adsorption and Reaction) approach22 to prepare core-shell QDs7,15,22 and core-multishell QDs23-26 (e.g., CdSe-CdS-(Cd,Zn)S-ZnS,23 CdS-CdSe-CdS24,26). In contrast to the standard SILAR technique, which is based on alternate injections of enough precursors to yield one monolayer of the cation or anion,22-26 in the present work the alternate addition of precursors is carried out dropwisely, thus making homogeneous nucleation even more unlikely. The low growth temperature further decreases the reactivity of the precursors, ensuring that heterogeneous nucleation and growth of CdSe at the surface of the CdTe nanocrystals prevails, and also prevents growth of the CdTe cores by Ostwald ripening18 or alloying due to inter-diffusion.27 Therefore, any change in the size (or shape) of the nanocrystals can be ascribed to the shell growth. As a representative example, Figure 1 shows TEM images of samples collected at different stages during the CdSe shell growth on spherical CdTe cores with a diameter of

40 Chapter 3

2.6 nm (5% size dispersion). It can be seen that the growth is highly anisotropic, initially yielding prolate heteronanocrystals (Figure 1, sample B, diameter: 2.7 nm; length: 3.9 nm, implying the growth of 0.66 nm CdSe “tips” in the length direction only), which evolve in trigonal pyramids (Figure 1, sample D, side length: 5.2 nm, implying a 1.6 nm CdSe shell under the assumption that the CdTe core occupies the center of the pyramids), and eventually undergo branching producing multipods (Figure 1, sample E; mostly bipods, but tripods, rods, matchsticks and more complex multibranched shapes are also present; branch diameter: 3 nm, branch length: 6-10 nm).

Figure 1. Overview TEM images of CdTe/CdSe heteronanocrystal samples B-E (see Table 1) collected at different stages during the CdSe shell growth. Insets show corresponding HRTEM images. The anisotropic growth can be understood considering the set of reaction conditions used in this work (viz., low reactivity precursors, low reaction temperature, slow precursor addition, coordinating solvent with higher affinity for Cd than for Se). Under these conditions the CdSe heteroepitaxial growth is slow and kinetically controlled (i.e., the rate limiting step is the CdSe formation at the surface, rather than the diffusion of

Luminescent CdTe/CdSe Colloidal Heteronanocrystals 41

precursors), and therefore the higher reactivity of the {000ī} Se terminated polar facets of the CdSe wurtzite structure will favor anisotropic growth along the c-direction.28-31 The growth evolution towards branched heterostructures, such as the CdTe/CdSe bipods and tripods shown in Figure 1 for sample E is probably due to the different crystal structures of the CdTe core (zincblende) and the growing CdSe shell (wurtzite), 10,28 as shown by powder XRD. Figure 2 shows that the CdTe core diffraction pattern matches the bulk lattice spacing of zinc blende CdTe (dashed lines). The diffraction pattern of the CdTe/CdSe heteronanocrystals (Sample E; Table 1) matches closely the wurtzite diffraction pattern of CdSe. This supports the formation of wurtzite CdSe on zinc blende CdTe. The preferred growth direction seems to be the (110) direction.

10 20 30 40 50 60

Inte

nsty

(a.u

.)

2θ

CdTe core

CdTe/CdSe Sample E

Figure 2. Powder X-ray diffraction of CdTe core sample, and CdTe/CdSe hetero structures (sample E). The positions of bulk reflections for zinc blende CdTe and wurtzite CdSe are shown as dashed lines. 3.3.2 Optical properties: absorption, photoluminescence, quantum yields and exciton lifetimes. Figures 3 and 4 show that the optical properties (viz., absorption, photoluminescence and exciton lifetimes) of the CdTe/CdSe heteronanocrystals change dramatically during the CdSe “shell” growth. The emission wavelength red shifts from 535 nm (2.32 eV) for the CdTe cores (Figure 3) to 770 nm (1.61 eV) for the CdTe/CdSe multipods (sample E in Figures 1,3,4 and Table 1). The absorption also red shifts but, as the dimension of the CdSe part of the heteronanocrystal increases, the absorption features

42 Chapter 3

are smeared out and the oscillator strengths at the emission energies become relatively smaller (Figure 3), while the RT photoluminescence decay times (i.e., the exciton lifetimes) become longer (Figure 4).

400 500 600 700 800

E

C

B

Nor

mal

ized

abs

orba

nce,

pho

tolu

min

esce

nce

Wavelength (nm)

Core

A

D

Figure 3. Absorption (solid lines) and photoluminescence (dashed lines) spectra of colloidal CdTe/CdSe heteronanocrystals with a 2.6 nm CdTe core and increasing CdSe shell dimension. Labels correspond to Table 1. Spectra are shifted vertically for clarity. The increase in the exciton lifetime observed for sample CdTe/CdSe B (18 ns, mono-exponential decay; Figure 4) with respect to the CdTe cores (multi-exponential decay, average lifetime: 8 ns; Figure 4) is mostly due to the decrease of the non-radiative decay rates, since it is accompanied by a substantial increase in the PL QYs (viz., from 7% to 46%, Table 1). Moreover, the radiative exciton lifetime for CdTe QDs of similar size (viz., 3.0 nm) has been reported to be 18 ns.32 Further growth of the CdSe part leads to increasingly longer exciton lifetimes and a change from single- to bi-exponential decay (viz., Sample C: 24 ns, 1-exp. decay; Sample E: 2-exp. decay, 60 and 220 ns, average lifetime: 100 ns). The faster component of the bi-exponential decay is attributed to CdTe/CdSe heteronanocrystals in which nonradiative exciton recombination is also relevant, while the slower component is attributed to purely radiative decay. It is worth noting that the featureless absorption tail observed for the CdTe/CdSe multipods (Figure 3, sample E) is at lower energy than the band edge absorption of both the CdTe core (see

Luminescent CdTe/CdSe Colloidal Heteronanocrystals 43

Figure 3) and the CdSe branches (565 nm for CdSe nanorods with 3.2 nm diameter and 11 nm length33-34). Therefore both the absorption tail and the photoluminescence peak can be unambiguously assigned to an indirect exciton.

0 100 200 300 400 500 600

101

102

103

104 Core A B C D E

PL

inte

nsity

(cou

nts)

Time (ns)

Increasing shell thickness

Figure 4. Photoluminescence decay curves of colloidal CdTe/CdSe heteronanocrystals with a 2.6 nm CdTe core and increasing CdSe shell dimensions (Labels are explained in Table 1.) The decay curve of the CdTe QDs used as cores is also shown (Core). These results indicate a progressive reduction of the electron-hole wave function overlap and the formation of an indirect exciton as the dimension of the CdSe part of the heteronanocrystal increases (reflected in the progressively longer lifetimes accompanied by a spectral red shift and the development of a featureless sub band gap absorption tail), ultimately leading to a characteristic type-II behavior3-6,9 for the CdTe/CdSe multipods (sample E). The gradual evolution of semiconductor heteronanocrystals towards a type-II optical behavior has been demonstrated for ZnSe/CdSe,4-5 CdTe/CdSe6-9 and CdS/ZnSe17 spherical core/shell heteronanocrystals, and observed to depend both on the core diameter and the shell thickness. This is consistent with the fact that the position of the energy levels of semiconductor nanostructures is strongly size-dependent,1,2 making the relative energy offsets and the degree of carrier localization in heteronanocrystals strongly sensitive to the dimensions (and shape) of both the core and the shell.9,17,35 Recent theoretical modeling of ZnSe/CdSe35 and CdS/ZnSe17core/shell nanocrystals has shown that the carrier localization regime becomes dependent on the core diameter and shell thickness in the case of small core diameters, evolving from type-I to type-II through an

44 Chapter 3

intermediate regime (“quasi-type-II”), in which one carrier is still delocalized over the entire volume of the heteronanocrystal. Calculations based on the effective mass approximation have also been carried out for a CdTe/CdSe core/shell system9 (core diameter: 3.9 nm) and show that the hole wave function is mostly localized in the core already for rather thin CdSe shells (0.2 nm), while the electron wave function remains delocalized through the whole core/shell QD up to a shell thickness of 0.9 nm, when it starts to localize in the shell. The transition from type-I to type-II behavior is thus not abrupt, leading to a gradual change in the optical properties as the localization of the electron and hole in different parts of the heteronanostructure becomes more pronounced, in good qualitative agreement with the results reported here. The use of the theoretical models above9,17,35 to quantitatively analyze our results is precluded by the fact that the heteronanocrystals are modeled in refs. 9, 17 and 35 as spherically symmetric core/shell nanostructures. Low PL QYs have been seen as an intrinsic limitation of type-II heteronanostructures, since the slower radiative recombination of indirect excitons should facilitate the dominance of nonradiative recombination.6,12 In this context the high PL quantum yields of the CdTe/CdSe heteronanocrystals investigated here (as high as 82% for prolate nanocrystals and 45% for multipods dispersed in chloroform, see Figure 3) are extremely remarkable, since they are in striking contrast with the extremely low PL QYs reported in the literature for CdTe/CdSe nanostructures, viz. ≤10% for core/shell QDs6,8,16 and ~0% for tetrapods,10 heteronanorods,11 and dumbbarbells.12 These high PL QYs indicate relatively slow nonradiative recombination rates, implying low defect concentrations and therefore attesting the high-quality of the CdTe/CdSe heteronanocrystals prepared in this work. The low defect concentration is probably due to the combination of slow heteroepitaxial growth and conditions that favor effective surface passivation, relaxation and reconstruction.18,19 An additional effect of the anisotropic growth is the smaller strain at the CdTe-CdSe heterointerfaces compared to spherical core/shell heteronanocrystals. The lattice mismatch between CdTe and CdSe (~6%) will lead to defects in thick shell spherical nanocrystals, but anisotropic growth will minimize the stress in the heterostructured nanocrystal and reduce the formation of non-radiative traps, since anisotropic heteronanocrystals, such as heteronanorods, can accommodate much larger lattice mismatches than concentric heterostructures.10 These results show that type-II heteronanocrystals can have high QYs, despite the longer radiative lifetimes of indirect excitons, provided the concentration of surface and interfacial defects is minimized.

Luminescent CdTe/CdSe Colloidal Heteronanocrystals 45

3.3.3 Temperature dependence of the optical properties Figure 5 presents the emission spectra of CdTe/CdSe heteronanocrystals (Figure 5a) and 2.6 nm CdTe QDs (Figure 5b) at several temperatures (293-383 K), and compares the temperature dependent photoluminescence quantum yields of CdTe QDs and a number of CdTe/CdSe heteronanocrystal samples dispersed in decalin (Figure 5c). It can be seen that the PL intensity of CdTe QDs is strongly quenched as the temperature increases, while that of the CdTe/CdSe heteronanocrystals remains essentially unaffected (or even shows a modest increase in some cases). The different behavior of CdTe QDs and CdTe/CdSe heteronanocrystals regarding thermally activated non-radiative relaxation is also reflected in the PL decay curves (Figure 6), which become faster and more non-exponential with increasing temperature for CdTe QDs (Figure 6a), consistent with an increasing contribution of non-radiative recombination to the exciton relaxation. In contrast, the PL decay curves of CdTe/CdSe heteronanocrystals (see Figure 6b for a representative example) remain (nearly)single-exponential and become slightly slower with increasing temperature, which correlates well with the absence of thermal quenching and the modest increase in the luminescence QYs observed upon increasing the temperature from 293 to 373 K (Figure 5). Temperature quenching of the photoluminescence of QDs (e.g. CdSe/ZnS core/shell) is a commonly observed phenomenon, both in colloidal suspensions or in solvent-free systems, such as QDs embedded in polymeric matrices and QD solids, and is ascribed to thermally activated carrier trapping and non-radiative recombination at defects.36-39 The behavior of the CdTe/CdSe heteronanocrystals is thus quite remarkable and unusual, indicating that the defect concentration in these nanocrystals is very low. The modest increase in the PL intensities accompanied by slightly longer exciton lifetimes suggest that the nonradiative recombination rates even decrease a little with increasing temperature. This behavior is reminiscent of the “luminescence temperature antiquenching” effect observed for CdSe QDs capped by linear alkylamines,40 and is probably also due to a reversible surfactant-assisted surface relaxation (and/or reconstruction). Figure 5 shows also that the PL peak red shifts as the temperature increases, for both the CdTe cores and CdTe/CdSe heteronanocrystals. The data presented in Figure 7 reveals that this thermally induced PL red shift is larger for the CdTe cores and decreases as the dimension of the CdSe part of the heteronanocrystal increases. The absorption spectra are also observed to red shift with increasing temperature, but a quantitative analysis is complicated by the fact that the lowest energy absorption peak also contains contributions from the lower energy tail of higher energy transitions (the bands are quite broad with respect to the energy separation between subsequent transitions). Since different absorption transitions may have different temperature dependences in terms of

46 Chapter 3

both energies and oscillator strengths, the maximum of the lowest absorption peak may not reliably reflect the temperature dependence of lowest energy exciton transition. The temperature dependent shift of the absorption and the emission spectra (as well as the accompanying PL intensity changes discussed above) are fully reversible, indicating that the samples do not undergo any chemical modification and that the temperature dependence of the optical properties can be thus ascribed to a band gap change.

450 500 550 600 6500

2

4

6

PL

inte

nsity

(cou

nts/

105 )

Wavelength (nm)

(a)

450 500 550 600 6500

2

4

6

8

10

PL

inte

nsity

(cou

nts/

10 4 )

Wavelength (nm)

(b)

290 300 310 320 330 340 350 360 370 380 390

20

40

60

80

100(c)

PL Q

uant

um y

ield

(%)

Temperature (K)

Core

A

B

C

Figure 5. Photoluminescence spectra of (a) CdTe/CdSe heteronanocrystals (Sample B, Table 1) at temperatures between 293 and 383 K, and (b) 2.6 nm CdTe QDs at temperatures between 293 and 363 K. (c) PL QY in decalin as a function of temperature for 2.6 nm CdTe QDs and prolate CdTe/CdSe samples (core, and samples A, B, and C as in Table 1).

Luminescent CdTe/CdSe Colloidal Heteronanocrystals 47

0 20 40 60 80101

102

103

104

Inte

nsity

(cou

nts)

Time (ns)

293 K 328 K 350 K

(a)

0 20 40 60 80101

102

103

104

Inte

nsity

(cou

nts)

Time (ns)

293 K 233 K 353 K

(b)

Figure 6. Photoluminescence decay of (a) CdTe QDs and (b) CdTe/CdSe QD (sample C) at different temperatures. The temperature dependence of the band gap (Eg) of bulk semiconductors is well described by the Varshni relation,41 which has been shown to be also valid for semiconductor nanocrystals36,42: Eg=E0-α⋅T 2/(T+β), where α is the temperature coefficient, β is approximately the Debye temperature of the material, and E0 is the band gap at 0 K. The values of α and β have been obtained for CdSe/ZnS nanocrystals [viz., (3.2±0.2)⋅10-4 eV/K and 220±30 K, ref. 36 and (4.2±0.5)⋅10-4 eV/K and 155±55 K, ref. 42; for α and β respectively], and are consistent with the values

48 Chapter 3

known for bulk CdSe [viz., (2.8-4.1)⋅10-4 eV/K and (181-315) K for α and β, respectively43]. However, α and β values for CdTe nanocrystals are not available in the literature, and the values reported for bulk CdTe are conflicting (e.g. α = -3.⋅10-4 eV/K in ref. 44; and 4.1⋅10-4 eV/K in ref. 45). Therefore, the α and β coefficients of CdTe nanocrystals were determined in this work, by fitting the Varshni relation to the temperature dependence of the PL peak positions of 2.6 nm CdTe QDs in the 3 to 280 K temperature range (Figure 8). The fit yields a band gap of 2.446 eV at 0 K, α = (5 ± 0.47)⋅10-4 eV/K and β = 102 ± 32 K, which is close to the Debye temperature (158 K) reported in literature.43

2.6 2.8 3.0 3.2 3.4

-0.08

-0.06

-0.04

-0.02

0.00

C

B

Core

A

PL

peak

pos

ition

shi

ft (e

V)

1000/T (K-1)

(a)

0.0 0.4 0.8

11

12

13

14

15

16

17

PL

shift

(meV

)

CdSe shell in length dimension (nm)

(b)

Figure 7. (a) The PL peak position as function of temperature for CdTe and CdTe/CdSe heteronanocrystals with increasing prolate shell thickness (samples A, B, and C, as defined in Table 1). (b) The total PL emission shift from 293 to 373 K for CdTe/CdSe heteronanocrystals with increasing prolate CdSe shell. These results imply that the band gap of CdTe depends more strongly on the temperature than that of CdSe, which explains the results reported above (Figure 7). As discussed

Luminescent CdTe/CdSe Colloidal Heteronanocrystals 49

above, the hole wave function is strongly localized in the CdTe core already for rather thin CdSe shells, while the electron wave function progressively localizes in the CdSe part of the heteronanocrystal as the CdSe dimensions increase. Consequently, the CdSe contribution to the PL energies becomes increasingly larger, leading to a smaller temperature dependent shift of the PL peak.

0 50 100 150 200 250 300

2.34

2.36

2.38

2.40

2.42

2.44

2.46

P

L pe

ak p

ositi

on (e

V)

Temperature (K)

Figure 8. Photoluminescence peak position of 2.6 nm CdTe QDs as a function of temperature. The solid line is the best-fit curve with the Varshni relation.

3.4 Conclusions In summary, the preparation methodology developed in this work yields highly efficient anisotropic CdTe/CdSe colloidal heteronanocrystals, with shapes tunable from prolate to branched heteronanocrystals. The high-quality and low defect concentration of these heteronanocrystals is attested by their high photoluminescence quantum yields and negligible thermally induced quenching up to temperatures as high as 373 K, which allowed the observation of reversible thermally induced spectral red shifts. A progressive reduction of the electron-hole wave function overlap and the formation of an indirect exciton are observed as the dimension of the CdSe part of the heteronanocrystal increases, ultimately leading to a characteristic type-II behavior for the CdTe/CdSe multipods. These results show that type-II heteronanostructures can have high photoluminescence quantum yields, which has important consequences for a number of potential applications (e.g. photovoltaic devices, lasers, LED’s).

50 Chapter 3

References 1 Alivisatos, A. P. J. Phys. Chem. 1996, 100, 13226 2 de Mello Donegá, C.; Liljeroth, P.; Vanmaekelbergh, D. Small 2005, 1, 1152 3 Hatami, F.; Grundmann, M.; Ledentsov, N. N.; Heinrichsdorff, F.; Heitz, R.;

Böhrer, J.; Bimberg, D.; Ruvimov, S. S.; Werner, P.; Ustinov, V. M.; Kopev, P. S.; Alferov, Zh. I. Phys. Rev. B 1998, 57, 4635

4 Balet, L. P.; Ivanov, S. A.; Piryatinski, A.; Achermann, M.; Klimov, V. I. Nano Lett. 2004, 4, 1485

5 Nanda, J.; Ivanov, S. A.; Htoon, H.; Bezel, I.; Piryatinski, A.; Tretiak, S.; Klimov, V. I. J. Appl. Phys. 2006, 99, 034309

6 Kim, S.; Fisher, B.; Eisler, H-J.; Bawendi, M. J. Am. Chem. Soc. 2003, 125, 11466 7 Li, J. J.; Tsay, J. M.; Michalet, X.; Weiss, S. Chem. Phys. 2005, 318, 82 8 Chou, P. T.; Chen, C. Y.; Cheng, C. T.; Pu, S. C.; Wu, K. C.; Cheng, Y. M.; Lai, C.

W.; Chou, Y. H.; Chiu, H. T. Chem. Phys. Chem. 2006, 7, 222 9 Oron, D.; Kazes, M.; Banin, U. Phys. Rev. B 2007, 75, 035330 10 Milliron, D.J.; Hughes, S. M.; Cui, Y.; Manna, L.; Li, J.; Wang, L.W.; Alivisatos,

A. P.; Nature 2004, 430, 190 11 Shieh, F.; Saunders, A. E.; Korgel, B. A. J. Phys. Chem. B 2005, 109, 8538 12 Halpert, J. E.; Porter, V. J.; Zimmer, J. P.; Bawendi, M. G. J. Am. Chem. Soc. 2006,

128, 12590 13 Chen, C. Y.; Cheng, C. T.; Yu, J. K.; Pu, S. C.; Cheng, Y. M.; Chou, P. T.; Chou,

Y. H.; Chiu, H. T. J. Phys. Chem. B 2004, 108, 10687 14 Chen, C. Y.; Cheng, C. T.; Lai, C. W.; Hu, Y. H.; Chou, P.T.; Chou, Y.H.; Chiu,

H.T. Small 2005, 1, 1215 15 Xie, R.; Zhong, X.; Basché, T. Adv. Mater. 2005, 17, 2741 16 Yu, K.; Zaman, B.; Romanova, S.; Wang, D.; Ripmeester, J. A. Small 2005, 1, 332 17 Ivanov, S. A.; Piryatinski, A.; Nanda, J.; Tretiak, S.; Zavadil, K. R.; Wallace, W.

O.; Werder, D.; Klimov, V. I. J. Am. Chem. Soc. 2007, 129, 11708 18 Wuister, S. F.; van Driel, F.; Meijerink, A. Phys. Chem. Chem. Phys. 2003, 5, 1253 19 de Mello Donegá, C.; Hickey, S. G.; Wuister, S. F.; Vanmaekelbergh, D.;

Meijerink, A. J. Phys. Chem. B 2003, 107, 489 20 Strickler, S. J.; Berg, R. A. J. Chem. Phys. 1962, 37, 814 21 Karstens, T.; Kobs, K. J. Phys. Chem. 1980, 84, 1871 22 Li, J. J.; Wang, Y. A.; Guo, W.; Keay, J. C.; Mishima, T. D.; Johnson, M. B.;

Peng, X. J. Am. Chem. Soc. 2003, 125, 12567 23 Xie, R.; Kolb, U.; Li, J.; Basché, T.; Mews, A. J. Am. Chem. Soc. 2005, 127, 7480 24 Battaglia, D.; Li, J. J.; Wang, Y.; Peng, X. Angew. Chem. Int. Ed. 2003, 42, 5035

Luminescent CdTe/CdSe Colloidal Heteronanocrystals 51

25 Battaglia, D.; Blackman, B.; Peng, X. J. Am. Chem. Soc. 2005, 127, 10889 26 Schill, A. W.; Gaddis, C. S.; Qian, W.; El-Sayed, M. A.; Cai, Y.; Milam, V. T.;

Sandhage, K. Nano Lett. 2006, 6, 1940 27 Bailey, R.E.; Nie, S. J. Am. Chem. Soc. 2003, 125, 7100 28 Yin, Y.; Alivisatos, A. P. Nature 2005, 437, 664 29 Peng, X. Adv. Mater. 2003, 15, 459 30 Rempel, J. Y.; Trout, B. L.; Bawendi, M. G.; Jensen, K. F. J. Phys. Chem. B 2005,

109, 19320 31 Talapin, D. V.; Koeppe, R.; Götzinger, S.; Kornowski, A.; Lupton, J.M.; Rogach,

A.L.; Benson, O.; Feldmann, J.; Weller, H. Nano Lett. 2003, 3, 1677 32 Wuister, S. F.; de Mello Donegá, C.; Meijerink, A. J. Chem. Phys. 2004, 121,

4310 33 Shabaev, A.; Efros, Al. L. Nano Lett. 2004, 4, 1821 34 Li, L.; Hu, J.; Yang, W.; Alivisatos, A. P. Nano Lett. 2001, 1, 349 35 Piryatinski, A.; Ivanov, S. A.; Tretiak, S.; Klimov, V. I. Nano Lett. 2007, 7, 108 36 Valerini, D.; Creti, A.; Lomascolo, M.; Manna, L.; Cingolani. R.; Anni, M. Phys.

Rev. B 2005, 71, 235409 37 Walker, G.W.; Sunder, V.C.; Rudzinski, C.M.; Wun, A.W. ; Bawendi, M.G. ;

Nocera, D.G. Appl. Phys. Lett. 2003, 83, 3555 38 Liu, T. –C.; Huang, Z. –L.; Wang, H. –Q.; Wang, J. –H.; Li, X. –Q.; Zhao, Y. –D.;

Luo, Q. –M. Anal. Chim. Acta 2006, 559, 120 39 Schöps, O.; Le Thomas, N.; Woggon, U.; Artemyev, M. V. J. Phys. Chem. B 2006,

110, 2074 40 Wuister, S. F.; de Mello Donegá, C.; Meijerink, A. J. Am. Chem. Soc. 2004, 126,

10397 41 Varshni, Y.P. Physica 1967, 34, 149 42 Joshi, A.; Narsingi, K. Y.; Manasreh, M. O.; Davis, E. A.; Weaver, B. D. Appl.

Phys. Lett. 2006, 89, 131907/1-3 43 Numerical Data and Functional Relationship in Science and Technology, edited by

K.H. Hellwege, Landolt-Börnstein, New Series, Group III, Vol. 17, Pt. B Springer-Verlag, Berlin, (1982)

44 Camassel, J.; Auvergne, D.; Mathieu, H.; Triboulet, R.; Marfaing, Y. Solid State Commun. 1973, 13, 63

45 Kosyachenko, L. A.; Sklyarchuk, V. M.; Sklyarchuk, Ye. F.; Ulyanitsky, K. S. Semicond. Sci. Technol. 1999, 14, 373

52 Chapter 3

53

Chapter 4

Cluster Synthesis of Branched CdTe Nanocrystals for Light-Emitting Diodes

Summary Highly luminescent cadmium telluride (CdTe) nanocrystals were synthesized using Li2[Cd4(SPh)10] as a reactive Cd cluster compound at relatively low temperature, making it a safe precursor for the large scale synthesis of CdTe nanocrystals. TEM showed that the shape of the CdTe nanocrystals changes from nanorods to branched structures with increasing reaction time. The nanocrystals show high luminescent quantum yields up to 37% for CdTe branched nanostructures, and as high as 52% for CdTe/CdS core shell heterostructures. CdTe/CdS nanocrystals were used to make light-emitting diodes in combination with organic layers for electron and hole injection. The devices show a maximum luminance efficiency of 0.35 cd/A.

54 Chapter 4 4.1 Introduction The excellent size-dependent, bright and narrow photoemission of colloidal semiconductor cadmium chalcogenide nanocrystals (NCs) combined with the flexibility in proccesability enables them to be used in solution processed hybrid organic-inorganic light-emitting diodes (LEDs). 1-8 Like CdSe, CdTe NCs are potentially attractive as emitter in light-emitting diodes, because of their bright and tunable luminescent properties. Compared to the frequently used CdSe core-shell particles,1-8 CdTe NCs may have an advantage in terms of band alignment. For CdSe NCs, the valence band level is at φ = -6.5 eV and, hence, there is a substantial (≥ 1 eV) injection barrier for hole injection from commonly used organic hole injection layers that have φ = -5.0 to 5.5 eV. The valence band level of CdTe NCs (φ = -5.5 eV) is much higher and provides a significant reduction of the barrier for hole injection compared to CdSe. At the same time, this causes a much higher sensitivity of the CdTe NCs towards oxidation under ambient conditions compared to CdSe. Few reports describe the use of CdTe as emitter in light-emitting devices. Electroluminescent devices based on CdTe NC and conducting polymers composites have shown an external quantum efficiency of 0.1%,9 while for layer-by-layer assembled multilayer devices of CdTe and poly(diallyldimethylammonium chloride) (PDDA) efficiencies of 0.51 % have been reported recently.10 The progress in the synthesis of cadmium chalcogenides has provided routes to high quality, monodisperse, crystalline nanoparticles and their corresponding core shell structures.11 A disadvantage in the preparation of CdS, CdSe, or CdTe NCs is the frequent use of dimethyl cadmium as cadmium source. More recent studies on the use of alternative precursors like cadmium oxide and cadmium salts (e.g. cadmium acetate) to replace the toxic and highly pyrophoric dimethyl cadmium have shown to produce high quality CdS, CdSe, and CdTe nanocrystals.12 These reactions rely on the injection of precursors at temperatures as high as 300 oC, due to the low reactivity of the precursors. These high reaction temperatures are not beneficial for the large scale production of nanocrystals. Milder synthesis routes in aqueous environment have been reported towards CdTe rods and dots, but these often make use of the highly toxic H2Te gas as precursor.13,14 While synthesis in aqueous media is advantageous for biological applications, it is less favorable for optoelectronic applications. In this paper we describe the use of a Cd precursor with a high reactivity that allows the formation of highly luminescent CdTe NCs under mild reaction conditions and we show that these CdTe NCs can be successfully applied to make LEDs. In order to provide an efficient and safe route to CdTe NCs that is free of organometallic cadmium precursors, we use a non-organometallic synthesis leading to highly luminescent rod

Cluster Synthesis of Branched CdTe NCs for Light-Emitting Diodes 55

shaped and branched CdTe NCs using Li2[Cd4(SPh)10] clusters15 as cadmium source in combination with elemental tellurium at 147 °C. Subsequent surface passivation of the CdTe NCs by growing a wider band gap CdS shell, leads to NCs exhibiting fluorescence quantum yields (QY) up to 52%. We analyze the CdTe nanoparrticles with transmission electron microscopy, X-ray diffraction, and (time-resolved) photoluminescence. The CdTe/CdS core shell NCs were used as the emissive layer in LEDs, fabricated using organic charge injection layers.

4.2 Experimental

Materials. Trioctylphosphine (TOP, tech. grade 90%), trioctylphosphine oxide (TOPO, 99%), hexadecylamine (HDA, 90%), dodecylamine (DDA, 98%), elemental sulfur (sublimated), tellurium (powder 30 mesh, 99.997%), and anhydrous chloroform, methanol, and toluene were purchased from Aldrich. LiNO3, Cd(NO3)2·4H2O, octadecylamine (ODA, >99%) and thiophenol were purchased from FLUKA. Cadmium(II) acetate (>99.99%) was purchased from Strem. Poly(3,4-ethylenedioxy-thiophene):poly(styrenesulfonate) (PEDOT:PSS) (EL grade) was obtained from HC Starck. All other chemicals were of analytical quality and were obtained from Sigma-Aldrich. The surfactants were degassed and dried at 120 °C for at least one hour under vacuum prior to use, all other chemicals were used as obtained. Li2[Cd4(SPh)10]. Li2[Cd4(SPh)10] clusters were synthesized using a modification of the method proposed by Dance et al.15 Cd(NO3)2·4H2O (10.5 g, 34 mmol) dissolved in methanol (30 mL) was added to a well-stirred mixture of thiophenol (9.3 g, 91 mmol) and triethylamine (12.7 ml, 91 mmol) in methanol (20 mL) at room temperature Subsequently LiNO3 (2.8 g, 40 mmol) in methanol (20 mL) was added and stirring was continued until all precipitate had dissolved. The Li2[Cd4(SPh)10] clusters were crystallized by cooling to -18 ºC. The obtained Li2[Cd4(SPh)10] crystals were filtered and dried under vacuum for 24 h. Typical CdTe synthesis. CdTe NCs were synthesized under a flow of argon with less than 1 ppm of water and oxygen. In a typical CdTe synthesis, Li2[Cd4(SPh)10] (0.32 g, 0.21 mmol) and elemental tellurium (0.09 g, 0.7 mmol) were mixed in a 50 ml flask containing DDA (5 g) and TOP (7 mL) at room temperature. The reaction mixture was slowly heated with an oil bath to 147 ºC under vigorous stirring. At 90 °C the reaction mixture turned light yellow as consequence of the elemental Te that started to dissolve in TOP. This was followed by a change in color to darker yellow above 100 °C (used as start) as a result of the nucleation of CdTe NCs. The crystal growth was followed by taking small aliquots at different time intervals that were quickly diluted and cooled in

56 Chapter 4 DDA/TOP at room temperature for optical characterization. The CdTe NCs were kept under protective inert atmosphere to prevent oxidation. CdTe synthesis with low Cd concentration. To study the effect of lower cadmium concentration was half the amount of Li2[Cd4(SPh)10] used (0.16 g, 0.102 mmol) and the original amount of elemental tellurium (0.09 g, 0.7 mmol). The other reaction conditions were kept the same as those described previously. CdS shell. In a regular synthesis of CdTe/CdS core-shell NCs, 4.5 mL of crude solution of CdTe NCs (58 mg) prepared as described above was diluted with chloroform (3 mL) and precipitated with methanol (10 mL). The NCs were collected by centrifugation. The collected NCs were dissolved in toluene (1 mL) and mixed with HDA (7.5 g) and TOPO (1.5 g) at 80 °C. The toluene was subsequently removed in 15 min. under vacuum at 100 °C. A solution of elemental sulfur (0.0272 g, 0.848 mmol) in TOP (2.5 mL) and a solution of cadmium acetate (0.20 g, 0.848 mmol) in TOP (2.5 mL) were alternatingly added dropwise during 60 min. to the reaction mixture at 165 ºC. Small aliquots at different stages of the CdS shell growth were quickly cooled in glass vials cooled on ice to quench the particle growth for further optical characterization. UV-vis absorption and fluorescence spectroscopy. Spectra were measured in dry and oxygen free toluene unless stated otherwise. UV-vis spectra were recorded using a Perkin–Elmer Lambda 900 spectrophotometer. Fluorescence spectra were measured with an Edinburgh Instruments FS920 spectrophotometer. Quantum yields were determined by comparing the integrated emission intensity with that of Rhodamine 101 in Ethanol containing 0.01% HCl (QY = 99%) on samples with an optical density <0.1 at the excitation wavelength of 520 nm. Time-correlated single photon counting (TCSPC) was used to determine the time resolved photoluminescence. The NC dispersions were excited at 400 nm using a picosecond diode laser (PicoQuant PDL 800B) operated at 2.5 MHz. The detector unit consisted of a Peltier-cooled Hamamatsu microchannel plate photomultiplier (R3809U-50). Transmission electron microscopy. HRTEM images were recorded using a TECNAI G2 20 transmission electron microscope (FEI Co., The Netherlands) operated at 200 kV. The samples for TEM were prepared by dipping an amorphous carbon copper (400-mesh) TEM grid into a chloroform solution of nanocrystals. The excess liquid was removed by a filtration paper. X-ray diffraction. XRD was recorded with a Rigaku D/Max B diffractometer (graphite monochromator, scintillation counter) using Fe Kα radiation (1.937 Å) operating at 40 kV, 30 mA. The data were collected from drop cast NCs on glass samples using a step scan mode with a step size of 0.02º and a counting time of 3 s per step in the range 2θ between 20° to 80º

Cluster Synthesis of Branched CdTe NCs for Light-Emitting Diodes 57

LED fabrication. The diodes were constructed on glass substrates with a patterned indium tin oxide (ITO) electrodes with a 53 nm thick layer of PEDOT:PSS to planarize the surface topology of the ITO layer and improve hole injection into the device. A 32 nm layer of polyvinylcarbazole (PVK) was spin coated onto the PEDOT layer, as a hole transport/electron blocking layer. The CdTe/CdS NCs were deposited by spin coating in inert atmosphere (< 1 ppm O2, < 1 ppm H2O) from a solution in cyclohexane (6 or 12 mg/ml), in which the PVK is not soluble. A 60 nm hole blocking/electron transport layer of 1,3,5-tris(2-N-phenylbenzimidazolyl)-benzene (TPBI) was thermally evaporated in high vacuum. The metal cathode for electron injection was obtained by successively evaporating 5 nm of barium and 100 nm of aluminum. Device characterization. The current-voltage (I-V) measurements were performed inside a glovebox using a Keithley 2400 source meter. The electroluminescence spectra were recorded with a fiberoptic Avantes AVS USB 2000 spectrometer. The light intensity was measured with a Hamamatsu S9219 photodiode connected to a Keithley 2400 source meter. The photodiode was calibrated using a Minolta LS110 luminance meter.

4.3 Results and Discussion 4.3.1 Synthesis and characterization The CdTe NCs were prepared by reacting Li2[Cd4(SPh)10] with elemental tellurium in a surfactant mixture of trioctylphosphine (TOP) and dodecylamine (DDA) at 147 °C under vigorous stirring. Monitoring the particle growth with UV-vis and PL spectroscopy (Figure 1) showed that during the first 30 min. the absorption and emission spectra of the CdTe NCs shift to longer wavelengths, corresponding to a rapid increase in particle size. Subsequently, the emission and absorption maxima shift more slowly and finally saturate. By replacing DDA by hexadecylamine (HDA) or octadecylamine (ODA) we found that the alkyl chain length of the surfactant has no significant effect on particle growth or on the optical properties. The particle diameter was estimated from the excitonic transition energy by comparing to experimental data from literature.12, 16 The emission maximum can be tuned between 500 and 613 nm by varying the reaction time from 2 to 180 min., corresponding to particle diameters of 2.5 to 3.8 nm (Figure 1).

58 Chapter 4

400 450 500 550 600 650 700

2.53 nm

3.85 nm

3.68 nm

3.53 nm

3.51 nm

3.41 nm

2.71 nm

3.04 nm

8 min

2 min

start

Wavelength (nm)

180 min

110 min

60 min

40 min

20 min

Figure 1. UV-vis and photoluminescence spectra of CdTe NCs monitored in time during growth in a typical synthesis at 147 ºC. Spectra were recorded at room temperature in DDA/TOP. The average confinement width in nm obtained from first absorption peak is indicated.

Cluster Synthesis of Branched CdTe NCs for Light-Emitting Diodes 59

During the first 20 min. two extra absorption peaks at ~425 and ~468 nm can be observed, well below the onset of the absorption. A lower concentration of Li2[Cd4(SPh)10] in the reaction leads to a significant increase of these distinct absorption peaks as shown in Figure 2. These peaks originate from very small CdTe clusters that disappear during the synthesis over the time scale of 1 hour, under the conditions used,17 in favor of larger particles as evidenced from the shifting absorption (Figure 2a) and emission bands (Figure 2b). It has been found that in the small size regime atomic clusters of certain sizes show a preferred thermodynamic stability. The shape and position of the peaks correspond to the spectra reported in the literature.17a, 18 The increased formation of these ultra small CdTe clusters at reduced Cd concentration is expected to be a consequence of the formation of very stable surface crystal facets that are mainly tellurium terminated. The relative large amount of thiophenol (present from the precursor) is expected to decrease the release of cadmium that can be consumed during CdTe formation.

400 500 6000.00

0.05

0.10

a

Abs

orba

nce

Wavelength (nm)

10 min

15 min

30 min

60 min

450 500 550 600 6500

3

5

8

10

b

Nor

mal

ized

PL

inte

nsity

Wavelength (nm)

10 min 15 min 30 min 60 min

Figure 2. (a) UV-vis spectra at different time intervals during CdTe growth at low Cd concentration at 147 °C showing the sharp absorption peaks at ≈ 435 nm and ≈ 480 nm. (b) Photoluminescence spectra (λex = 400 nm), normalized to the emission maximum, revealing a narrow PL emission at different time intervals during CdTe growth.

The full width at half maximum of the luminescence band remains about 30 nm during particle growth, irrespective of the initial Li2[Cd4(SPh)10] concentration used (Figure 1 and Figure 2b). A sharp emission peak often corresponds to a narrow particle size distribution (<10 %), which is usually accompanied by a sharp peak in the absorption spectrum.11a In contrast, the UV-vis spectra (Figure 1 and Figure 2a) of the CdTe NCs show a steep onset of the absorption, but no well-defined peak. To resolve this issue, transmission electron microscopy (TEM) was used to investigate the evolution of particle shape during the synthesis. The TEM graphs revealed

60 Chapter 4 the formation of rod shaped NCs from 20 min. onwards. Figure 3a shows a TEM image of rod shaped NCs obtained after 1 h as an example.

Figure 3. (a) TEM images of: (a) Rod shaped CdTe NCs after 1 h growth. (b) Branched CdTe NCs obtained after 3 h growth. (c) Rod shaped and branched CdTe/CdS NCs after shell growth. The nanorods do not have a constant width and are sometimes bended, explaining the absence of a well-defined absorption peak. The excitation energy in the nanorods will localize at the site with the largest diameter (corresponding to the smallest band gap). Hence, the narrow photoluminescence and steep absorption onset can be explained by the fact that the thickest part of the NCs is relatively monodisperse (3.8 nm after 3 h). Previous work on the synthesis of CdTe NCs in an alkylamine/TOP surfactant mixture using dimethyl cadmium resulted in the formation of spherical particles.19 Since the synthetic procedures used here are further identical, we infer that thiophenolate originating from the cadmium clusters may act as a coordinating surfactant that causes an anisotropic particle growth. Accordingly, the presence of thiophenolate on the surface of purified particles was confirmed with FT-IR spectroscopy by peaks (3054, 1643, 1576, 1081 and 1026 cm-1) that can only correspond to stretching vibrations of the thiophenolate (not shown). Together with the increase in nanorod diameter over time, a significant increase in fluorescence QY is observed, from <1% at 2 min. to as high as 37 % after 3 h. This is a remarkably high QY considering the elongated and branched structure of the NCs, which reveal a large surface area. After 3 h the emission wavelength and QY stabilize at 613 nm. TEM revealed that branched structures were formed after heating for 3 h (Figure 3b). X-ray diffraction on powder samples of the particles after 3 h heating (Figure 4) reveals diffraction peaks that match those of the zinc blende crystal structure of CdTe.16b

a b c

Cluster Synthesis of Branched CdTe NCs for Light-Emitting Diodes 61

20 30 40 50 60 70 80

Inte

nsity

(a.u

.)

2θ

(111)

(220)

(311)

Figure 4. Powder X-ray (λ = 1.936 Å) diffraction pattern of CdTe NCs formed after 2 h reaction time in a HDA/TOP surfactant mixture. To further improve the photoluminescence QY of the CdTe NCs, the surface was passivated with a wider band gap CdS shell. CdS shell growth was performed using cadmium acetate and elemental sulfur in a mixture of HDA and trioctylphosphine oxide (TOPO) at 165 ºC. The growth of the CdS shell causes a 30 nm red shift in both absorption and emission (Figure 5). The fluorescence QY of the core CdTe NCs (synthesized in DDA/TOP, 3 h. reaction time) in the HDA/TOPO reaction mixture of the shell synthesis is at 1.8%, which is much lower than the original 37% in DDA/TOP. As a result of the washing step between the CdTe core synthesis and the shell growth a certain amount of the weakly bound surfactants was removed, resulting in nonradiative surface traps. The development of the fluorescence QY during the shell synthesis shows an immediate strong rise, because of the decrease of CdTe surface traps. A maximum QY of 52% was reached for the branched CdTe/CdS NCs. Figure 3c shows the TEM image of branched CdTe/CdS core shell particles, with cores (synthesized in DDA/TOP, 3 h.). The spherical parts on the ends of the rod shaped and branched particles indicate that the CdS growth mainly takes place on the ends of the rods.

62 Chapter 4

400 450 500 550 600 650 700

1.8%

start

24%

39%

0.2 mmol

0.4mmol

40%

0.6 mmol

0.8 mmol46%

0.85 mmol 52%

Wavelength (nm)

Figure 5. UV-vis and photoluminescence spectra of CdTe/CdS core shell NCs monitored during CdS shell growth at 165 ºC upon adding different amounts of precursor. Spectra were recorded at room temperature in chloroform solution. Figure 6 compares the time resolved luminescence of the CdTe core particles during the synthesis for a sample taken after 8 min. and after 180 min. Generally, the luminescence lifetime of NCs shows a multi-exponential decay with short components attributed to the luminescence from core states and long components attributed to luminescence from surface states. 20 The decays in Figure 6 were fitted with a triple exponential decay. The resulting lifetimes and the corresponding abundance of the lifetimes are summarized in Table 1. The relative increase of the longer components during the CdTe synthesis indicates a decrease in surface traps during the synthesis and a corresponding increase in the amount of luminescence from surface states compared to luminescence from core states. A similar decrease in surface traps can be seen during the

Cluster Synthesis of Branched CdTe NCs for Light-Emitting Diodes 63

CdS shell synthesis. Purification of the nanocrystals gave a large reduction of the contribution of the longer components of the luminescence lifetime, which shows that surface defects were created. This is an indication that the amine ligands bind weakly to the surface of the nanocrystals. The long components in the lifetime increase again after the CdS shell synthesis, which shows that this is a good way to decrease surface traps.

0 25 50 75 100 125 1500.1

1

10

100

final

a

PL

inte

nsity

(cou

nts/

100)

Time (ns)

180 min8 min

0 20 40 60 80 100 120 1400.1

1

10

100b

PL

inte

nsity

(cou

nts/

100)

Time (ns)

CdTe core purified

CdTe/CdS core shell

Figure 6. Luminescence decay curves (λex = 400 nm) measured at the emission maximum for the CdTe samples. (a) Taken after 8, 180 min, and at the end (final) of the CdTe core synthesis. (b) Purified CdTe NCs at the start of the CdS shell synthesis and the final CdTe/CdS core shell particles at the end of the synthesis of the CdS shell. Table 1. Luminescence (λex = 400 nm) lifetimes obtained after fitting the luminescence decays of Figure 6 with a triple exponential decay for three CdTe NCs samples taken during synthesis of the core (after 8, 180 min, and at the end) and two samples taken during synthesis of the CdS shell (the purified CdTe core at the start and the final CdTe/CdS core-shell).Values in parenthesis refer to the individual contributions of each lifetime component to the total decay.

Sample τ1 (ns) τ2 (ns) τ3 (ns) CdTe core 8 min 1.6 (39 %) 4.3 (59 %) 15.9 (2 %) CdTe core 180 min 1.6 (3 %) 5.7 (61 %) 16.0 (36 %) CdTe core final 1.4 (16 %) 5.5 (59 %) 16.1 (25 %) CdTe core purified 0.7 (35 %) 2.8 (42 %) 10. 2 (23 %) CdTe/CdS core-shell 1.6 (24 %) 12.3 (62 %) 32.2 (14 %)

64 Chapter 4 4.3.2 Electroluminescence The high PL QY and narrow emission peak of the CdTe/CdS NCs make them an attractive candidate as an emitter in LEDs. Most of the work concerning LEDs containing inorganic semiconductor NCs, is based on CdSe and CdSe/CdX-ZnX core shell NCs.1-8 CdSe, however, has a valance band at φ = -6.5 eV, which is well below that of most organic materials (φ = -5.0 to -5.5 eV) used for hole injection. Recent work showed the electroluminescence from thiol/polymer-capped CdTe NCs, in a layer-by-layer assembled device where multiple layers of capped NCs were deposited alternating with PDDA between an ITO anode an aluminum cathode.10c This EL device showed a luminescence efficiency of 0.4 cd/A and a brightness peaking at 1.42 cd/m2.10c These devices did not show diode behavior, but confirm that CdTe is an attractive material in EL devices. We constructed LEDs using CdTe NCs in a hybrid polymer-NC diode structure, similar to what has been reported for CdSe NCs.4 The layout of the device and the corresponding band diagram of the used materials are shown in Figure 7. The higher energy of the valance band of CdTe at -5.5 eV compared to CdSe at -6.5 eV results in a smaller barrier for hole injection from the highest occupied molecular orbital of PVK, which could lead to higher device performance in hybrid NC-LEDs. To make the LEDs, a thin layer of CdTe/CdS NCs was spin coated from cyclohexane onto a 32 nm PVK layer. To enhance electron injection a thermally evaporated layer of TPBI (60 nm) was deposited onto the NC layer. The TPBI layer acts as an electron transport and hole blocking layer, and reduces exciton quenching by the metal cathode by increasing the distance between the emissive layer and the metal cathode.21

Cluster Synthesis of Branched CdTe NCs for Light-Emitting Diodes 65

ITO PEDOTPVK CdS CdTe CdS TBPI Ba Al-8-7-6-5-4-3-2-10

b

3.94.1 4.1

2.71.8

4.3

2.5

5.5

6.6 6.6 6.7

5.45.055.0Ene

rgy

(eV

)

Figure 7. (a) Schematic structure of the device. (b) Energy levels of various materials, in eV with respect to vacuum.

The electroluminescence spectrum of CdTe/CdS NC LEDs (Figure 8) peaks at 623 nm due to CdTe/CdSe NC emission. The peak in the EL spectrum is slightly red shifted (10 nm) compared to the PL spectrum of the same NCs in solution. This small red shift is a consequence of interparticle energy transfer from smaller particle to larger particles when the NCs are in close proximity. An additional emission between 400 and 500 nm is observed, which is attributed to PVK emission. The PVK emission can be reduced by increasing layer thickness of the CdTe to yield virtually pure NC emission, but at the cost of luminance efficiency (Figure 8).

a

66 Chapter 4

400 500 600 700 800 9000

100

200

300

400

500

EL

inte

nsity

(a.u

.)

Wavelength (nm)

Figure 8. Electroluminescence (EL) spectra of a CdTe/CdS NC-LED operated at 42 mA/cm2. The closed squares represent the EL of a thin NC layer device (spin coated from 6 mg/ml solution), the open circles represent a thicker layer of similar CdTe/CdS NCs (spin coated from 12 mg/ml solution). Different batches of NCs were, used explaining the small shift in NC emission. The luminance, current density, and luminance efficiency are plotted in Figure 9. The device shows a clear diode behavior, current and electroluminescence are only induced under a forward bias at around 14 V. This relatively high onset voltage is in the same range as observed for similar CdSe/CdS NC devices4 and reveals that the anticipated beneficial effect of the alignment of the levels of PVK and CdTe NCs is not yet reflected in the device characteristics. We assume that it is mainly caused by the resistance of the NC layers and the presence of the organic TPBI layer, which is used to confine excitons and migrating holes in the emissive layer. The maximum efficiency of this CdTe/CdS NC LED was 0.35 cd/A at a 16 V. This efficiency is about half of that of a similar device using CdSe/CdS (0.65 cd/A)4 and slightly less than the electroluminescent (0.4 cd/A) device10 based on CdTe NCs alternating with PDDA layers. Compared to the recently reported CdSe based NC-LEDs, however, where luminance efficiencies ranging from 1 to 2.8 cd/A were reached,6-8 the performance is lower. In our view, the device efficiency is limited by the resistance of the NC layer, caused by the presence of the insulating bound and non bound surfactants (HDA/TOPO). Extra washing steps of the NCs as described in the Experimental section reduce the amount of surfactants, but resulted in a dramatic decrease in luminescence quantum yield of the NCs. This decrease can be explained by considering that most of the passivating CdS shell material is formed at the ends of the branches (Figure 3c) and that washing will reduce the surface bound surfactants that passivate surface defects on parts of the branches where little CdS has been deposited.

Cluster Synthesis of Branched CdTe NCs for Light-Emitting Diodes 67

0 5 10 15 200

2

4

6

8

10

12

0

1

2

3

4

5

Lum

inan

ce in

tens

ity (c

d/m

2 )

Voltage (V)

Current density (m

A/cm

2)

a

0 1 2 3 4 50.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40b

Lum

inan

ce e

ffici

ency

(cd/

A)

Current density (mA/cm2)

Figure 9. Electroluminescence characteristics of a CdTe/CdS NC-LED (spin coated from 6 mg/ml solution) (a) Luminance and current density as function of the of the applied bias (b) Luminance efficiency versus current density.

4.4 Conclusion The new synthesis of CdTe NCs using Li2[Cd4(SPh)10] as Cd cluster compound and elemental Te represents a safe and reproducible route for the formation of highly luminescent rod shaped and branched particles at relatively low temperatures. Branched CdTe core and CdTe/CdS core-shell NCs with fluorescence quantum yield of 37 and 52 %, respectively were obtained. The reaction conditions are the same as for the procedure

68 Chapter 4 that uses dimethylcadmium.19 This illustrates the high reactivity of the cluster compound. Since Li2[Cd4(SPh)10] clusters can easily be synthesized in multi gram scale, the route presents an easy and safe alternative for the large scale synthesis of CdTe NCs.

The fluorescence quantum yields luminescent properties of the branched CdTe NC cores (37%) and CdTe/CdS NCs (52%) are remarkably high compared to values previously reported for branched NCs, where the maximum quantum yield is below 10%.22 These quantum yields also compare well to earlier reported quantum yields for the best spherical CdTe QDs made in water or by organometallic approach, that reach 40% 14 and 65%,19 respectively. Application of the CdTe/CdSe NCs in an LED resulted emission with an luminance efficiency of 0.35 Cd/A, which is in the range found for devices based on highly luminescent CdSe NCs. This work showed that it is possible to use CdTe/CdS NCs as an alternative for CdSe NCs in LEDs. Further work is necessary to establish the true scope of these emitting particles for NC-LEDs.

References 1 Colvin, V. L.; Schlamp, M. C.; Allivisatos, A. P. Nature 1994, 370, 354 2 Dabboussi, B. O.; Bawendi, M. G.; Onitsuka, O.; Rubner, M. F. Appl. Phys. Lett. 1995, 66, 1316 3 Coe, S.; Woo, W. K.; Bawendi, M. G.; Bulović, V. Nature 2002, 420, 800 4 Hikmet, R. A. M.; Chin, P. T. K.; Talapin, D. V.; Weller, H. Adv. Mater. 2005, 17, 1436 5 Steckel, J. S.; Snee, P.; Coe-Sullivan, S.; Zimmer, J. P.; Halpert, J. E.; Anikeeva, P.; Kim, L. A.; Bulović, V.; Bawendi, M. G. Angew. Chem. Int. Ed. 2006, 45, 5796 6 Li, Y.; Rizzo, A.; Cingolani, R.; Gigli, G. Adv. Mater. 2006, 18, 2545 7 Niu, Y.-H.; Munro, A. M.; Cheng, Y.-J.; Tian, Y.; Liu, M. S.; Zhao, J.; Bardecker, J. A.; Jen-La Plante, I.; Ginger, D. S.; Jen, A. K.-Y. Adv. Mater. 2007, 19, 3371 8 Sun, Q.; Wang, Y. A.; Li, L. S.; Wang, D.; Zhu, T.; Xu, J.; Yang, C.; Li, Y. Nature Photonics 2007, 1, 717 9 (a) Gaponik, N. P.; Talapin, D. V.; Rogach, A. L. Phys. Chem. Chem. Phys. 1999, 1, 1787 (b) Gaponik, N. P.; Talapin, D. V.; Rogach, A. L.; Eychmüller, A J Mater. Chem. 2000, 10, 2163 10 (a) Gao, M.; Lesser, C.; Kirstein, S.; Möhwald, H.; Rogach, A.; Weller, H. J. Appl. Phys. 2000, 87, 2297 (b) Chen, W.; Grouquist, D.; Roark, J.

Cluster Synthesis of Branched CdTe NCs for Light-Emitting Diodes 69

J. Nanosci. Nanotech. 2002, 2, 47 (c) Bertoni, C.; Gallardo, D.; Dunn, S.; Gaponik, N.; Eychmüller, A. Appl. Phys. Lett. 2007, 90, 034107/1-3 11 (a) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993, 115, 8706 (b) Talapin, D. V.; Rogach, A. L.; Kornowski, A.; Haase, M.; Weller, H. Nano Lett. 2001, 1, 207 12 (a) Peng, Z. A.; Peng, X. Nano Lett. 2001, 1, 333 (b) Talapin, D. V.; Mekis, I.; Götzinger, S.; Kornowski, A.; Benson, O.; Weller, H. J. Phys. Chem. B 2004, 108, 18826 (c) Yu, W. W.; Qu, L.; Guo W.; Peng, X Chem. Mater. 2003, 15, 2854 13 Deng, D. W.; Qin, Y. B.; Yang, X.; Yu, J. S.; Pan, Y. J. Cryst. Growth 2006, 296, 141 14 Gaponik, N; Talapin, D V; Rogach, A L; Hoppe, K; Shevchenko, E V; Kornowski, A; Eychmüller, A; Weller, H J. Phys. Chem. B 2002, 106, 7177 15 Dance, I. G.; Choy, A.; Scudder, M. L. J. Am. Chem. Soc. 1984, 106, 6285 16 (a) Yu, W. W.; Wang, A. Y.; Peng, X. Chem. Mater. 2003, 15, 4300 (b) Rajh, T; Micic, O. I.; Nozik, A. J. J. Phys. Chem. 1993, 97, 11999 17 (a) Rogach, A. L.; Katsikas, L.; Kornowski, A.; Su, D.; Eychmüller, A.; Weller, H. Ber. Bunsen Ges. Phys. Chem. 1997, 101, 1668 (b) Herron, N.; Calabrese, J. C.; Farneth, W. E.; Wang, Y. Science 1993, 259, 1426 18 Wuister, S. F.; Driel, F.; Meijerink, A. J. Luminescence 2003, 102, 327 19 Talapin, D. V.; Haubold, S.; Rogach, A. L.; Kornowski, A.; Haase, M.; Weller, H. J. Phys. Chem. B 2001, 105, 2260 20 Wang, X.; Qu, L.; Zhang, J.; Peng, X.; Xiao, M. Nano Lett. 2003, 3, 1103 21 Era, M.; Adachi, C.; Tsutsui, T.; Saito, S. Chem. Phys. Lett. 1991, 178, 488 22 Mohamed, M. B.; Tonti, D.; Salman, A. A.; Chergui, M. ChemPhysChem. 2005, 6, 2505

70 Chapter 4

71

Chapter 5

Polarized Light-Emitting Quantum Rod Diodes

Summary Polarized light-emitting quantum rod diodes were fabricated using macroscopically oriented quantum rods with an aspect ratio of 2.5 as the emissive layer. Devices were constructed by sandwiching a rubbed layer of quantum rods between two organic layers with electron and hole conducting properties. The LED emitted red light at 620 nm with a luminescence efficiency of 0.65 cd/A, and external a quantum efficiency of 0.49%. Electroluminescence from the diodes showed polarized emission and the intensity of the light polarized in the direction parallel to the long axis of the rods was 1.5 times higher than in the direction perpendicular to it.

72 Chapter 5 5.1 Introduction Colloidally synthesized semiconductor nanocrystals (NCs), also referred to as “quantum dots”,are receiving ever more attention due to their fascinating properties, which differ substantially different from those of their respective bulk materials.1 In such nanometer sized particles, a gradual transition from bulk to molecular structure occurs as the particle size decreases. The particles show size-dependant optical and electronic properties. For example, the band gaps of these materials may increase by several electron volts compared to that of the bulk materials with decreasing particle size.2 This is reflected in the absorption and the photoluminescence spectra of the materials, which shift by hundreds of nanometers with decreasing particle size. Most of the theoretical work done so far has concentrated on explaining the behavior of quantum dots.3 Progress in colloidal synthesis made it possible to prepare highly crystalline and fairly monodisperse nanocrystals with controllable particle size.4,5 Thanks to the possibility to have an outstanding degree of control over the particle size and monodispersity, it is also possible to control the shape of particles. Rod-shaped particles of several nanometers in radius and with various aspect ratios can now be routinely produced.6,7 Such quantum rods also show narrow-band, linearly polarized emission,8,9,10 and anisotropic charge-transport properties. They can be partially aligned in stretched polymer films6 and can form liquid-crystalline phases in concentrated dispersions.11,12 The alignment of quantum rods on a macroscopic scale may make it possible to prepare a novel class of artificial solids with strong, intrinsically anisotropic optical and electric properties. The possibility of aligning CdSe quantum rods, for example, would imply a major advantage for the further development of hybrid solar cells13 and photovoltaic devices.14 The tunable sharp emission characteristics of quantum dots also makes them interesting for use as chromophores in light-emitting diodes.15,16,17,18 Such dots have been used in various configurations to produce light-emitting diodes. Although their performance lags behind that of organic light-emitting diodes, their potential stability and sharp emission characteristics make them, nevertheless, attractive for use in light-emitting diodes. The polarized photoluminescence of quantum rods is used in this study to produce polarized electroluminescence in a light-emitting diode.

Polarized Light-Emitting Quantum Rod Diodes 73

5.2 Experimental

Synthesis. The synthesis and optical properties of the quantum rods used and single-rod photoemission characteristics can be found in a previous publication.19 Sample preparation and optical spectroscopy. In order to orient quantum rods in a polymer film, a solution of ultrahigh-molecular-weight polyethylene in decalin containing quantum rods was produced. Cooling of the solution led the formation of a gel. Subsequent drying and stretching of this gel to 100 times its original length resulted in highly oriented polyethylene film containing quantum rods.

A Perkin-Elmer (Labda 900) UV-vis spectrometer was employed and polarized light was used to investigate polarization dependence of the absorption properties of the oriented quantum rods. The emission and excitation spectra were measured using a standard Perkin-Elmer (LS50B) luminescence spectrometer fitted with polarizers. Unpolarized light at 420 nm was used to excite the quantum rods and the emitted intensity was measured for two polarization directions. In the same way, excitation spectra were measured for two orthogonal polarization directions for light emitted at 623 nm. A randomly oriented sample was used as a reference, and corrections were made regarding any effects concerning measurement configuration and polarization dependence of gratings of the instrument. Diode construction. The diodes (3 mm × 3 mm) were constructed on glass substrates coated with a patterned indium tin oxide (ITO) electrode and a 100 nm thick layer of poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) to improve the surface topology of the ITO layer without any risk of shorts. The PEDOT layer was subsequently coated with a layer of polyvinylcarbazole (PVK), which is known to have hole-transport / electron-blocking properties. Quantum rods were then deposited on the PVK layer from a solution in cyclohexane, in which PVK is not soluble. After inducing macroscopic orientation of the quantum rods by rubbing with a velvet cloth, an electron transport/hole blocking layer of 1,3,5-tris(N-phenylbenzimidazol-2-yl)benzene (TPBI) 20 was deposited in high vacuum using physical vapor deposition. Finally, barium and aluminum were successively evaporated as cathode.

74 Chapter 5 5.3 Results and Discussion 5.3.1 Polarized Photoluminescence In this study, core-shell quantum rods were used with a cadmium selenide core and a cadmium sulfide shell, passivated by trioctylphosphine (TOP), hexadecylamine (HDA), and trioctylphosphine oxide (TOPO) capping. An electron micrograph of the rods used in the study is shown in Figure 1.

Figure 1. Transmission electron microscopy (TEM) image of the CdSe/CdS quantum rods. The rods had a diameter of about 4.5 nm and an average aspect ratio of 2.3. The photoluminescence quantum efficiency of the rods in chloroform was estimated to be 50%. We used two different methods to study the polarization-dependent spectral properties of the rods. Figure 2a, b show the absorption and photoemission spectra obtained, for polarization directions parallel and perpendicular to the stretching direction, for rods in stretched ultrahigh-molecular-weight polyethylene (UHMWPE) films with excitation at 420 nm. Figure 2b also shows the polarized excitation spectrum obtained for emission at 623 nm.

Polarized Light-Emitting Quantum Rod Diodes 75

300 400 500 600 7000.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Abso

rban

ce

Wavelength

|

\\

(a)

400 450 500 550 600 650 7000

20

40

60

80

100

120

Inte

nsity

(a.u

.)

Wavelength (nm)

|

//

//

|

(b)

Figure 2. a) Absorbance b) Photoluminescence excitation (dotted lines) and emission (solid lines) as function of wavelength, for rods in a stretched UHMWPE film for polarization parallel (//) and perpendicular (⊥ )to the stretching direction.

It can be seen that the rods oriented in the film show high absorption and also high photo- emission/excitation intensities in the polarization direction parallel to the stretching axis. The results obtained here are in good agreement with an earlier study,19 in which single-rod polarized photoluminescence measurements were performed on similar rods and photo-emission was measured as a function of angle between the polarization direction and the long axis of the rod. The quantity So, which is related to the order parameter, S, describes the degree of orientation of the rods as So= S/X =(A//-A⊥)/(A//+2A⊥), was estimated to be 0.13, on the basis of the absorbance in the direction parallel (A//) and perpendicular (A⊥) to the long axis. The parameter, X, is given by (1-1.5

76 Chapter 5 sin2θ)-1, where θ is the angle between the absorption transition moment and the long axis of the molecules.21 This method has been successfully used in the past to align dichoric dye molecules to an order parameter of almost unity. If assumed that the rods are also perfectly oriented, we arrive at X =7.6. We estimated the ratio of the intensity of the light polarized in the direction of orientation (I//) to the intensity of the light polarized in the orthogonal direction (I⊥) in photoemission, R=( I/// I⊥) to be 2.5. Having established the maximum degree of polarized emission that is to be expected from well oriented quantum rods, a method for inducing macroscopic orientation of the rods was studied, which would also be applicable in the production of light-emitting diodes. To this end, the quantum rods were spin-coated on a substrate, to obtain a layer with a thickness of about 70 nm. The sample was subsequently heated to about 130 oC, and rubbed uniaxially using a velvet cloth. During the rubbing, the thickness of the layer is decreased to 30 nm. The degree of orientation of the rods in rubbed layers was characterized using polarization-dependent absorption and photoemission measurements. Using the above equation, the order parameter, S, of the rods estimated from the polarized absorbance measurements, was found to be ~ 0.9. The rubbed layers also showed polarized emission and the intensity of light polarized in the direction of rubbing was estimated to be 2.2 times higher than the intensity of the light polarized in lateral directions. This method was highly reproducible and delivered oriented layers of quantum rods. 5.3.2 Polarized Electroluminescence We also tried to produce devices showing stable emission at reasonably high intensities. Most of the work previously done on light-emitting diodes containing quantum dots led to rather low efficiencies because of a mismatch of energy levels and charge unbalance. Coe et al.22 showed that when a mono-layer of quantum dots is formed between hole-transporting/ electron-blocking and electron-transporting/hole-blocking layers, the performance of the light-emitting quantum-dot diodes can be significantly improved. These authors, however, relied on a phase separation process during which a monolayer of quantum dots was formed between two organic layers with hole- and electron-conducting properties. This enabled them to produce light-emitting diodes that showed emission with an external quantum efficiency of 0.52%. The diode structure presented here consists of various layers in order to make reproducible devices with relatively stable emission characteristics. A schematic representation of the diode and the energy levels with respect to vacuum are shown in Figures 3a,b.

Polarized Light-Emitting Quantum Rod Diodes 77

Al Ba TPBI CdSe/CdS PVK PEDOT ITO Glass

V

ITO PEDOT PVK CdS CdSe TPBI Ba Al-8

-7

-6

-5

-4

-3

-2

-1

0(b)

2.5

4.1 4.4

2.81.9

4.3

6.6 6.5 6.25.4

4.75.0Ener

gy (e

V) LUMO

HOMO

Figure 3. a) Structure of the device b) Energy levels of various materials, in electron volts, with respect to vacuum. TPBI: 1,3,5-tris(N-phenylbenzimidazol-2-yl) benzene; PVK: polyvinyl-carbazole; PEDOT: Poly(3,4-ethylenedioxythiophene); ITO: indium tin oxide; LUMO: Lowest Unoccupied Molecular Orbital; HOMO: Highest Occupied Molecular Orbital. The layered structure described in the Experimental section and in Figure 3a resulted in devices that could be made in a reproducible way. We also tried to optimize the thickness of the layers in order to realize a high efficiency, while minimizing emission from other layers in the device. The best result was obtained with a 62 nm thick PVK layer and a TPBI layer with a thickness of 60 nm. Figures 4a,b show electro-optical properties of a light-emitting quantum-rod diode. Figure 4a shows the electroluminescence obtained for two orthogonal polarization directions. It can be seen that the polarized light emission in the direction of rubbing was higher than that in the direction perpendicular to it. The ratio of the intensities for orthogonal polarization directions was estimated to be 1.6, which is rather less than the value of 2.2 observed for the photoluminescence spectra. This lower value may be associated with the fact that not all the rods in the layer were properly oriented during rubbing. Quantum rods at the bottom interface (the interface with PVK) will most probably be oriented to a lesser extent. The device’s emission spectrum shows a blue emission, corresponding to the emission from the PVK. This indeed indicates that recombination process took place largely around the PVK interface. Also observable in

(a)

78 Chapter 5 this figure is that the PVK emission is not polarized, which means that, as expected, the PVK is not oriented, and that only the quantum rods are aligned, resulting in the observed polarized-light emission.

400 500 600 7000.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

400 420 440 460 4800.000

0.002

0.004

0.006

EL in

tens

ity (a

.u.)

Wavelength (nm)

//

|

(a)

EL

inte

nsity

(a.u

.) Wavelength (nm)

0.1 1 10 100 10000

100

200

300

400

500

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 10 20 300

1x103

2x103

3x103

Lum

inan

ce (c

d/m

2 )

Current density (A/m2)

Lum

inan

ce e

ffici

ency

(cd/

A)

(b)

Cur

rent

den

sity

(A/m

2 )

Bias (V)

Figure 4. a) Electroluminescence (EL) spectrum from a diode for a polarization parallel (//) and perpendicular (⊥) to the rubbing direction. b) Luminance and luminance efficiency as function of the current density without a polarizer. The inset shows the current density as a function of the applied bias voltage.

Polarized Light-Emitting Quantum Rod Diodes 79

The electrical properties of the device can be seen in Figure 4b, in which the luminance and luminance efficiency, obtained without using a polarizer, have been plotted as function of the current density. The inset shows the current density as a function of the voltage. The devices work as a diode and light emission is induced only when ITO is positively biased at around 11 V. This high onset voltage is due to the presence of the organic layers, which are necessary to confine recombination excitons in the quantum-rod layer and to avoid exciton quenching (by TPBI) at the metal cathode. The efficiency of the device at an emission of 450 cd/m2 was estimated to be 0.65 cd/A, corresponding to an external quantum efficiency of 0.49%.

5.4 Conclusion In summary, we have introduced a method for obtaining polarized-light-emitting diodes using quantum rods. Polarized-light emission is of great industrial significance, as it can be used in backlights for displays. Polarizers are currently used in displays based on light emitting diodes to obtain improved daytime contrast, causing 50% absorption of the generated light. Light losses can be avoided by using polarized-light-emitting diodes in displays. However, the polarization efficiency will have to be significantly improved to make such devices suitable for such applications. In our research, we used rods with rather a small aspect ratio. In literature,23 however, it has been demonstrated that rods with a very high aspect ratios can be used to obtain photoluminescence in which the light polarized along the long axis of the rods has an intensity that is 50 times that of the light polarized in lateral directions. Such rods with high aspect ratios are also assumed to yield electroluminescence with high polarization ratios. We have demonstrated that such quantum-rod devices can be obtained with greater efficiency in a reproducible way. Quantum rod/dot devices offer the advantages of higher stability, narrow emission-band characteristics, and high efficiency, through mixed spin-triplet and spin-singlet exciton characteristics, and we believe the results presented here do not present the ultimate performance feasible with these materials.

References 1 Alivisatos, A. P. Science 1996, 271, 933 2 Bawendi, M. G.; Carroll, P. J.; Wilson, W. L.; Brus, L. E. J. Chem. Phys. 1992, 96, 946

80 Chapter 5 3 Brus L. E. J. Chem. Phys. 1983, 79, 5566 and 1984, 80,4403 4 Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Annu. Rev. Mater. Sci. 2000, 30, 545 5 Kim, F.; Song, J.; Yang, P. J. Am. Chem. Soc. 2002, 124, 14316 6 Peng, X.; Manna, L.; Yang, W. D.; Wickham, J.; Scher, E.; Kadavanich, A.; Alivisatos, A. P. Nature 2000, 404, 59 7 Li, L. S.; Hu, J.; Yang, W.; Alivisatos, A. P. Nano Lett. 2001, 1, 349 8 Hu, J.; Li,. L. S.; Yang, W.; Manna, L.; Wang, L. W.; Alivisatos, A. P. Science 2001, 292, 2060 9 Elfros, A. L. Phys. Rev. B 1992, 46, 7448 10 Kan, S.; Makari, T.; Rothenberg, E.; Banin, U. Nature Mater. 2003, 2, 155 11 Li, L. S.; Walda, J.; Manna, L.; Alivisatos, A. P. Nano Lett. 2002, 2, 557 12 Li, L. S.; Alivisatos, A. P. Adv. Mater. 2003, 15, 408 13 Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002, 295, 2425 14 Scher, E. C.; Manna, L.; Alivisatos A. P. Phil.Trans. R. Soc. London A 2003, 361, 241 15 Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994, 370, 354 16 Mattoussi H.; Radzilowski, L. H.; Dabbousi, B. O.; Fogg, D.E.; Schrock, R.R.; Thomas, E. L.; Rubner, M. F.; Bavendi. M. G. J. Appl. Phys. 1999, 86, 4390 17 Gao, M.; Lesser, C.; Kirstein, S.; Mohwald, H.; Rogach, A. L.; Weller, H. J. Appl. Phys. 2000, 87, 2297 18 Schlamp, M. C.; Peng, X.; Alivisatos, A. P. J. Appl. Phys. 1997, 82, 5837 19 Talapin, D. V.; Koeppe, R.; Götzinger, S.; Kornowski, A.; Lupton, J. M.; Rogach, A.; Benson, L. O.; Feldmann, J.; Weller, H. Nano Lett. 2003, 3, 1677 20 Anthopoulos, T. D.; Markham, J. P. J.; Namdas, E. B.; Samuel I. D. W.; Lo, S. C.; Burn, P. L. Appl. Phys. Lett. 2003, 82, 4824 21 Fiksinski, K.; Bauman, D.; Skibinski, A.; Stolarski, R. Dyes Pigm. 1991, 15, 203 22 Coe, S.; Woo, W. K.; Bawendi, M.; Bulovic, V. Nature 2002, 420, 800 23 Wang, J.; Gudiksen, M. S.; Duan, X.; Cui, Y.; Lieber, C. M. Science 2001, 293, 1455

81

Chapter 6

Energy Transfer and Polarized Emission in Cadmium Selenide Nanocrystal Solids with

Mixed Dimensionality

Summary Excited state energy transfer from spherical green-emitting nanocrystals as donor to rod shaped red-emitting nanocrystals as acceptor is shown to result in a predominant red fluorescence in mixed films. For this purpose red-emitting core shell CdSe/CdS quantum rods with high fluorescence quantum yield of ηF = 45% were synthesized and used to create mixtures with green-emitting core shell CdSe/CdS quantum dots (ηF = 40%). For this donor-acceptor combination the Förster distance is less than 6.6 nm and close to sum of the diameters of the dots and rods. Hence, only quantum dots directly neighboring a quantum rod can efficiently participate in the energy transfer. A simple rubbing technique was used to uniaxially align the quantum rods dispersed in thin.

82 Chapter 6 6.1 Introduction Spherical nanosized semiconductor crystals exhibit size-dependent electronic and optical properties.1 As a result of quantum confinement of the wave functions for holes and electrons, the absorption and fluorescence spectra of these quantum dots move to shorter wavelengths with decreasing particle size.2 Monodisperse quantum dots with narrow, tunable emission wavelengths and high fluorescence quantum efficiencies have been prepared3,4 and have been used in molecular diagnostics as fluorescent markers,5-8 in light-emitting diodes,9,10 or in solar cells.11 In addition to regulating the size, also the shape of the nanocrystals can be controlled. High quality, nearly monodisperse rod-shaped nanocrystals with different aspect ratios have been prepared.12-15 Nanorods show linearly polarized absorption and fluorescence along the long axis of the crystal.12,16-20 Interestingly, these rod-shaped crystals can be macroscopically aligned by dispersion and subsequent stretching in polymer films,12,21 or by rubbing a layer with a velvet cloth.21 In this way new composites with intrinsically anisotropic optical and electrical properties have been created.21 Polarized light sources are desirable in applications such as backlights for liquid crystal displays. Such light sources can have a significant energy saving advantage over conventional backlight systems where absorbing polarizers that discard at least 50% of the incident light intensity are used. In such an application one can make use of a lamp emitting unpolarized light and use a polarized light emitter for converting the unpolarized light into polarized light. This method is well established for luminescent molecules or conjugated polymers oriented in films or liquid crystalline hosts. However, polarized light emitting systems also tend to show polarization dependent absorption which again leads to loss of light. Therefore one needs to remove the polarization dependence of absorption from such a system. This has first been demonstrated by using randomly orientated sensitizer molecules that collect the incident light by isotropic absorption and then efficiently transfer the energy to a uniaxially orientated photoluminescent polymer, from which colored light with a high degree of linear polarization is emitted.22 Here we explore the same concept using blends of semiconductor nanocrystals with mixed dimensionality rather than molecular systems. We demonstrate that polarized light can be created by transferring input energy that is isotropically absorbed by spherical green-emitting CdSe/CdS core-shell quantum dots to nearby oriented red-emitting CdSe/CdS core-shell quantum rods via Förster energy transfer. Future application of these mixed nanocrystal solids may be found in placing such a composite material directly on top of a light-emitting diode (LED) chip to obtain a polarized light source.

Energy Transfer and Polarized Emission in CdSe NC Solids 83

Förster energy transfer between differently sized quantum dots23-26 and between quantum dots and dyes or polymers has been observed.27,28 In general, Förster energy transfer requires spectral overlap between the emission spectrum of the donor and the absorption spectrum of the acceptor. To ensure efficient energy transfer and subsequent emission of light, both the donor and the acceptor should possess high fluorescent quantum yields. Unfortunately, the fluorescence quantum yield of quantum rods12,29 is often significantly less than those of spherical quantum dots, where quantum yields above 60% are not uncommon. The larger surface area of the rods is considered to be an important reason for the reduced quantum yields. Passivation of the surface by creating a shell with a wider band gap has been successfully applied to quantum rods30,31 and CdSe quantum rods with a ZnS or CdS shell exhibiting fluorescence quantum yields up to 40% have been reported.29,32 The dilemma to overcome is that thick passivating shells can increase the fluorescence quantum yield but at the same time increase the distance between the cores of the dots and rods. This reduces the energy transfer efficiency that strongly depends (∝ R-6) on the distance (R) between donor and acceptor. We demonstrate that by applying a thin CdS shell on CdSe rods, the fluorescence quantum yield can be maintained at sufficiently high level to ensure dot to rod excitation energy transfer. In mixed dot-rod films in which the rods are macroscopically oriented this gives rise to essentially unpolarized absorption and significantly polarized emission.

6.2 Experimental Materials. Trioctylphosphine (TOP, tech. grade 90%), trioctylphosphine oxide (TOPO, 99%), hexadecylamine (HDA, 90%), cis-trans decalin (99%), elemental sulfur (sublimated), selenium powder 200 mesh (99.99%) and anhydrous chloroform and toluene were purchased from Aldrich. Cadmium acetate (+99.99+%) was purchased from Chempur. n-Tetradecylphosphonic acid (TDPA) was obtained from Alfa Aesar. All reagents were used as received, except TOPO and HDA. These reagents were dried and degassed prior to use by heating under vacuum, at 120 °C for 1.5 h. CdSe/CdS quantum rod synthesis. CdSe quantum rods were synthesized under argon flow according a slight modification of the method published by Bunge et. al.15 Selenium (0.06 g, 0.76 mmol) dissolved in TOP (2.4 g) and toluene (0.1 g) was swiftly injected into a reaction mixture of TDPA (0.22 g), TOPO (3.77 g), and cadmium acetate (0.103 g, 0.45 mmol) at 340 ºC. After stirring 7 min. at 300 ºC, toluene (7 mL) was injected into the reaction mixture, which rapidly reduces the temperature below the nucleation and growth temperatures of the crystals. Subsequently, the reaction mixture was further diluted using toluene (10 mL). To separate the CdSe quantum rods from excess precursors and

84 Chapter 6 surfactants, the quantum rods were precipitated by adding methanol (20 mL), followed by centrifugation. The precipitate was dissolved in chloroform (4.5 mL). For shell synthesis, the purified CdSe quantum rods (~15 mg in 1.5 mL chloroform) were redispersed in a mixture of TOPO (16 wt%) and HDA (83 wt%) at 60 ºC. Subsequently, the chloroform was removed in vacuum. Elemental sulfur (0.062 g, 1.94 mmol) and cadmium acetate (0.19 g, 0,83 mmol) dissolved in TOP (4 mL) were added dropwise during 50 min at 130-140 °C, followed by an annealing step of 60 min. at 100 ºC. The rod shaped crystals were subsequently isolated by dissolving the reaction mixture in chloroform (10 mL), followed by precipitated by addition of methanol (20 mL). The rods were isolated by centrifugation, dried overnight in vacuum, and subsequently stored under nitrogen. CdSe/CdS quantum dot synthesis. Spherical CdSe quantum dots were synthesized under a flow of argon in a mixture of HDA (4 g) and TOPO (8 g).37 A solution consisting of cadmium acetate (0.103 g, 0.45 mmol) and elemental selenium (0.15 g, 1.90 mmol) dissolved in TOP (4 mL) was swiftly injected into the HDA/TOPO mixture at 270 ºC under vigorous stirring. As soon as the color changed to yellow-orange, the nanocrystal growth was stopped by cooling the reaction mixture by injection of toluene (10 mL). The solution was diluted with toluene (10 mL). The nanocrystals were isolated by precipitation via addition of methanol (20 mL), followed by centrifugation. The isolated quantum dot crystals (50 mg) were dissolved toluene (2 mL) and added to a mixture of TOPO (1.5 g) and HDA (7.5 g). Subsequently, elemental sulfur (0.021 g, 0.65 mmol) and cadmium acetate (0.07 g, 0.30 mmol) dissolved in TOP (4 mL) were added dropwise during 30 min. to the reaction mixture at 147 ºC, to grow a monolayer shell, followed by annealing for 45 min. at 90 ºC. To separate the dots from the excess amount of surfactants and non-reacted precursors, the crude reaction mixture was dissolved in chloroform (15 mL), followed by precipitation by addition of dry methanol (20 mL) and isolation using centrifugation. Optical spectroscopy. UV-vis spectra were recorded using a Perkin–Elmer Lambda 900 spectrophotometer. Fluorescence spectra were measured with an Edinburgh Instruments FS920 spectrophotometer. Linear polarization of the luminescence was measured using the FS920 or alternatively using a home-built system employing a photoelastic modulator and collecting the luminescence in line with the direction of the depolarized excitation beam, in order to avoid artifacts resulting from photoselection.33 Time resolved photoluminescence was measured by time-correlated single photon counting (TCSPC) using a 400 nm picosecond diode laser (PicoQuant PDL 800B) operated at 2.5 MHz. The detector consisted of a Peltier-cooled Hamamatsu microchannel plate photomultiplier (R3809U-50). Spectra were measured in chloroform solution or as solid films on glass substrates. Photoluminescence quantum yields were estimated using fluorescein (0.1 M NaOH, ηF = 93%)34 and rhodamine 101 (in ethanol + 0.01% HCl, ηF = 100%).35 All solutions had an optical density <0.1 at the excitation wavelength (520 nm) to minimize

Energy Transfer and Polarized Emission in CdSe NC Solids 85

re-absorption and avoid absorbance saturation. The quantum yield was derived from luminescence spectra, by correcting for the optical density and the refractive index of the solvents used for sample and reference.36 Transmission electron microscopy. TEM images were recorded using TECNAI G2 20 (FEI Company, The Netherlands) operated at 200 kV. The TEM samples were prepared by dipping an amorphous carbon film (400 mesh) on a copper grid into the nanocrystal dispersion in chloroform. The excess liquid was removed with a filtration paper. Molar absorption coefficient. The molar absorption coefficient of the nanocrystals was determined using a combination of TGA, TEM, and UV-Vis spectroscopy on once-purified nanocrystals. In thermogravimetric analysis a sample of nanocrystal solid (2-5 mg) was heated under nitrogen while recording the weight as function of the temperature. A steep decrease in weight with temperature occurs in the range where the surfactants evaporate or decompose, followed by a much more gentle decrease. From the inflection point separating the two regions, the weight percentage of CdSe/CdS in the singly-purified samples can be determined. From the average particle size, as determined using TEM, it is possible to estimate the weight of a single particle, using the bulk densities of CdSe and CdS. In combination with the UV-vis absorbance of known concentrations of singly-purified nanocrystals in chloroform it is then possible to determine the absorption coefficient per mole particle. We note that using TGA, there is no uncertainty with respect to the presence of bound or unbound organic ligands provided that the same batch of nanocrystals is used in the entire procedure. The results obtained (see text) are consistent with those reported by Peng et al. (Ref. 39) who used atomic absorption to determine the Cd concentration. Sample preparation. The mixed rod:dot solid films were prepared after an additional washing step by mixing solutions of the purified nanocrystals in the desired ratios, followed by spin coating on clean glass substrates. Aligned films were prepared by spin coating the mixed rod:dot dispersions from chloroform on glass substrates coated with a 100 nm thick layer poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT:PSS). The orientation of the rods was induced by rubbing the film with a piece of velvet cloth.

6.3 Results and Discussion CdSe quantum rods were synthesized from cadmium acetate and selenium in a mixture of TOP, TOPO, and TDPA following a procedure similar to the method of Bunge et. al.15 Quantum rods collected from the reaction mixture exhibited fluorescence that maximizes at 585 nm with a fluorescence quantum yield (ηF) of about 8%. TEM images

86 Chapter 6 revealed an average width of 2.7 ± 0.3 nm and length of 9.8 ± 1.5 nm. To increase the fluorescence quantum yield we chose to passivate the surface states of the rods by growing a shell of CdS. To obtain a thin layer, shell growth was performed at a relatively low temperature (130-140 °C) in the presence of cadmium acetate and sulfur, followed by annealing at 100 °C. This procedure resulted in highly crystalline quantum rods with a length of about 10.5 ± 1.8 nm and a diameter of about 2.9 ± 0.4 nm (Figure 1a and b).

Figure 1. HRTEM images of (a, b) CdSe/CdS quantum rods and (c, d) a mixture of CdSe/CdS quantum dots and quantum rods. The passivated CdSe/CdS rods show an excitonic peak at 570 nm in the absorption spectrum (Figure 2, solid line) and are highly fluorescent (ηF = 45%) with an emission maximum at 584 nm. By comparing the sizes before and after shell growth, it is clear that the passivation procedure has only a minor effect on the dimensions of the rods. In addition, the fluorescence maximum remains almost constant. To rationalize the 5-fold improvement in fluorescence quantum yield we assume that the low growth temperature results in limited CdS surface coverage. Together with the effect of annealing, however, this is appears a sufficient and convenient method to produce thin shell quantum rods exhibiting high fluorescence quantum yields.

Energy Transfer and Polarized Emission in CdSe NC Solids 87

300 400 500 600 700

dots

Nor

mai

lzed

abs

orba

nce

Wavelength (nm)

rods

Norm

alized fluorescence

Figure 2. Normalized absorption and fluorescence spectra (λexc = 350 nm) of CdSe/CdS quantum dots (dashed lines) and of CdSe/CdS quantum rods (solid lines) in chloroform solution at room temperature. Spherical CdSe/CdS quantum dots were synthesized from cadmium acetate and selenium in a mixture of TOP, TOPO, and HDA analogous to the method developed by Talapin.37 The CdSe core had a diameter of 2.2 nm, as determined using the first absorbance peak at 482 nm in reference with literature values.30,38,39 The CdSe cores were passivated by growing a shell of CdS using cadmium acetate and sulfur following a similar procedure as used for the rods. TEM images of the CdSe/CdS quantum dots showed an average diameter of 3.0 ± 0.4 nm, implying that a CdS shell of about 0.4 nm was formed on the core particles. The final CdSe/CdS quantum dots exhibit a first absorption peak at 504 nm with a corresponding fluorescence maximum at 524 nm and a quantum yield of ηF = 40% . Figures 1c and 1d show TEM images of a mixture of dots and rods drop cast from chloroform on a TEM grid. Figure 2 reveals that there is a good overlap between the emission spectrum of the quantum dots and absorption spectrum of the quantum rods, while there is only minor overlap between their emission spectra. These properties make the combination very suitable for the intended energy transfer. We studied the energy transfer by producing various mixtures of quantum dots and rods in solution. Due to large distance between the particles in a solution little or no energy transfer is to be expected and the spectrum (Figure 3) serves as a reference indicating the relative emission intensities of dots and rods with direct excitation. By removing the solvent, the distance between the particles decreases dramatically, resulting in a strong change in the relative intensity of the fluorescence bands in favor of the emission of the quantum rods (Figure 3). The fact that the weak emission of the rods in

88 Chapter 6 the solution becomes dominant in a thin film indicates that in the solid film dots and rods come in close enough proximity to enable excited state energy transfer. This leads to a reduction of the quantum dot emission and increased emission originating from the rods.

500 550 600 6500

5

10

0

5

10

15 Solution

Fluo

resc

ence

inte

nsity

(cou

nts/

104 )

Wavelength (nm)

Solid

Fluorescence intensity (counts/104)

Figure 3. Fluorescence spectra (λexc = 350 nm) of mixture of quantum rods and dots in a 1:30 ratio (by number) in chloroform solution (solid circles) and as a solid spin-coated film on glass (open circles). At the excitation wavelength the optical density was less than 0.1. Time-resolved fluorescence was used to investigate the energy transfer in more detail. Figure 4a shows the fluorescence decay traces recorded at 530 nm (corresponding to the quantum dot emission) for pure dots (in solution and film) and for dots in solid mixtures with rods. Figure 4b displays the corresponding fluorescence dynamics at 590 nm for pure quantum rods (in solution and film) and in the mixtures. The fluorescence lifetime of the pure nanocrystals (both dots and rods) decreases in solid films as compared to their solutions. This decrease is a consequence of a change in surface passivation in the films. Also interparticle energy transfer in the pure nanocrystal solids tends to reduce the fluorescence intensity and lifetime via exciton quenching at trap sites. When comparing the time dependent fluorescence of dot-rod mixtures and pure dots in solid films, a small additional decrease in lifetime is observed (Figure 4a). In contrast, the lifetime of the rod emission increases with increasing fraction of dots in the mixtures. These observations are in good agreement with earlier reports where a similar behavior was observed for energy transfer from smaller to larger nanocrystals.24 The decrease in fluorescence lifetime of the dots upon addition of rods supports the proposed energy transfer mechanism. It is of interest to note that the emission lifetime of the rods increases at lower rod:dot ratios. This supports the view that in the pure quantum rods layers, the emission lifetime is limited by interparticle energy transfer. The fluorescence lifetime traces of both pure and mixed

Energy Transfer and Polarized Emission in CdSe NC Solids 89

films are clearly multi-exponential and the energy transfer is expected to take place with different rate constants due to the inhomogeneous nature of the films.

0 20 40 60 80

102

103

104

0 20 40 60 80

dots solution dots film rot:dot (1:65) rot:dot (1:3) IRF

Fluo

resc

ence

inte

nsity

(cou

nts)

Time (ns)

Emission at 530 nm Emission at 590 nm

rods solution rod:dot (1:65) rod:dot (1:3) rods film

(a) (b)

Figure 4. Time-resolved fluorescence intensity of (a) the green-emitting quantum dot energy donor, and (b) the red-emitting quantum rod energy acceptor in solution and as solids in pure and mixed films. The rod:dot ratios are by number. The instrument response function (IRF) is shown in panel (a). To determine the length scale for the energy transfer, the Förster distance was determined from the absorption and fluorescence spectra using the overlap integral J defined as,40

λλλελ dFJ A4

0D )()(∫

∞

= (1)

where FD is the normalized donor emission, εA the molar absorption coefficient of the acceptor, and λ the wavelength. The Förster distance is defined as the distance at which energy transfer efficiency is 50%. This Förster radius is proportional to the overlap integral according to,40

6/1

F42

0 ])D([211.0 JnR ηκ −= (2)

where κ2 accounts for the relative orientation of donor and acceptor transition dipole moments. For colloidal dots the rotationally averaged orientation factor has recently been shown to be 2/3,41 a value that also applies for a quantum dot donor and long-axis

90 Chapter 6 polarized nano-rod acceptor transition. ηF(D) is the luminescence quantum yield of the donor in the absence of acceptor, and n is the refractive index of the solvent. The molar absorption coefficients at 350 nm were determined using thermogravimetric analysis (TGA) together with TEM and UV-vis spectroscopy to be 1.2 × 106 Lmol-1cm-1 for the rods and 6.1 × 104 Lmol-1cm-1 for the dots. For the dots these values are in reasonable agreement with the empirical function relating εA to the diameter.39 Using the absorption and emission spectra for the nanocrystals in solution (Figure 2), the overlap integral J was found to be 2.3 × 1016 Lmol-1cm-1nm4. Using this value together with ηF(D) = 0.40 and n =1.45, the Förster distance is R0 = 6.6 nm in chloroform. It is important to note that this value is likely an upper limit for the actual value in the solid films because there the fluorescence quantum yield in films is lower than in solution. Also the higher refractive index of CdSe (n = 2.506) compared to chloroform would reduce R0 to ~ 4.6 nm (Equation 2). Considering that the core shell rods and dots have diameters of 2.9 ± 0.4 nm and 3.0 ± 0.4 nm, respectively, only the quantum dots in direct vicinity of the rods are expected to contribute to the energy transfer. The fluorescence of mixed films was recorded for different rod:dot ratios (Figure 5a). The fluorescence intensity of the dots decreases sharply with increasing number of rods. Similar changes have been observed in mixtures of differently sized quantum dots.42 At a rod:dot ratio of 1:30, the remaining emission of the dots is reduced to ~50%. Taking the average dimensions as determined by TEM and assuming that the dots can align along the rods in a hexagonal close packing, indeed approximately 30-40 dots can simultaneously be in direct contact with the rods. Hence the 50% quenching at this concentration favorably agrees with the Förster distance of 6.6 nm, being close to the sum of the diameters of the particles. Increasing the concentration of rods leads to a further quenching. Complete quenching of the dot fluorescence, however, is only observed at a 1:1 ratio. At such concentrations one dot is likely in the vicinity of several rods, increasing the probability for energy transfer. The emission of the rods already maximizes at relatively low concentrations (rods:dots = 1:30) and remains fairly constant until it decreases for high rod concentrations (1:1) to a level similar of that of pure rods. The strong fluorescence of the rods at low concentration is consistent with the proposed excited state energy transfer. The decrease in red emission at higher rod concentrations is tentatively ascribed to energy transfer between rods, a process that will contribute to quenching at trap sites. The experimental fluorescence intensity of the rods and dots versus their relative concentration in the thin films is compared in Figure 5b to the intensity that would be expected if no energy transfer would occur. For the latter curve we estimated the fluorescence intensity from the relative concentrations and their absorption coefficients. Figure 5b reveals that the increase in fluorescence intensity of the rods at low

Energy Transfer and Polarized Emission in CdSe NC Solids 91

concentration cannot be explained by direct excitation. We conclude that it is a direct consequence of energy transfer from dots to rods and further enhanced by avoiding energy transfer amongst neighboring rods at low concentrations. The optical density of the films at wavelengths corresponding the luminescence of the dots, is too low (OD < 0.02) for energy transfer via the trivial mechanism involving emission and reabsorption of light.

500 550 600 6500

4

8

12

16

20

24

Fluo

resc

ence

inte

nsity

(cou

nts/

104 )

Wavelength (nm)

rods dots 1:30 rod:dot 1:14 rod:dot 1:3 rod:dot 1:1 rod:dot

a

0.0 0.2 0.4 0.6 0.8 1.00

4

8

12

16

20

241.0 0.8 0.6 0.4 0.2 0.0

Fluo

resc

ence

inte

nsity

(cou

nts/

104 )

Relative concentration rods

Relative concentration dots

b

Figure 5. (a) Fluorescence spectra for dots and rods in thin films at different concentrations (ratios are by number). The spectra are corrected for total absorption at the excitation wavelength (350 nm). (b) Fluorescence intensity for dots (circles) and rods (squares) as function of the relative concentration in mixed solids. Solid markers represent the experimental data (from panel a) and the open markers the expected fluorescence intensity for direct excitation assuming that energy transfer does not occur.

92 Chapter 6 As a result of energy transfer, the absorption features of the dots are expected to be present in the excitation spectrum of the rod emission. Figure 6 shows indeed a difference in the excitation spectra recorded at 610 nm for pure rods and for rods in a mixture with dots. Compared to the spectrum of the pure rods there is a distinct additional component in the region corresponding to the absorption of dots. By taking the difference of the normalized excitation spectra relative to the excitation spectrum of the mixture, it is possible to estimate the fraction of the emission at 610 nm originating from a photoexcitation on quantum rod created via energy transfer from a quantum dot. This fraction plotted versus wavelength (Figure 6, open triangles) resembles the absorption spectrum of the dots.

350 400 450 500 550 6000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0

20

40

60

80

100

pure rods rod:dot (1:10)

Nor

mai

lzed

PL

inte

nsity

at 6

10 n

m

Wavelength (nm)

Fraction energy transfer (%)

Figure 6. Excitation spectra at 610 nm of a film of pure rods and of a mixed rod:dot film (ratio 1:10 by number). The spectra are normalized at the first excitonic peak at 570 nm. The open triangles represent the fraction of energy transfer from the dots to rods and resembles the absorption spectrum of the dots. To investigate polarized emission from mixed films, the rod shaped nanocrystals were macroscopically oriented by rubbing a thin layer of the material deposited on a glass substrate covered with PEDOT:PSS using a velvet cloth. This results in linearly polarized emission from the rod shaped crystals with a degree of polarization p = 0.25. Here p is defined as: p = (I|| − I⊥)/(I|| + I⊥), with I|| and I⊥ are the emission intensities parallel and perpendicular to the orientation direction respectively. Under the same sample preparation, the polarization of the emission of the dots is practically zero (p < 0.005). For mixed films, the luminescence from the aligned rods is also polarized (p = 0.27). In Figure 7, luminescence spectra for the mixed film (rod:dot = 1:14, by number)

Energy Transfer and Polarized Emission in CdSe NC Solids 93

are illustrated. The spectra were corrected for the polarization sensitivity of the monochromators and detector, and have been normalized to the unit maximum emission intensity for the dots, which was shown to be essentially unpolarized. The photons emitted by the rod shaped crystals are predominantly polarized parallel to the orientation direction of the rods induced by the rubbing. However in contrast to the behavior observed for pure rods, the aligned rods mixed with dots do not show a strong dependence on the polarization direction of the light used for exciting the system. This indicates that the light used for exciting the system is mainly absorbed by the dots and then transferred to the rods which show polarized emission.

500 550 600 6500

10

20

30

40

50

60 EM || EX || EM || EX ⊥ EM ⊥ EX || EM ⊥ EX ⊥

Nor

mal

ized

PL

Inte

nsity

(cou

nts/

104 )

Wavelength (nm)

QD emission

QR emission

p = 0.27

Figure 7. Polarized fluorescence spectra from a rubbed film containing a rod:dot mixture (1:14 by number) measured parallel and perpendicular to the rubbing direction. The spectra were recorded with a polarized excitation source at 350 nm. We also studied time resolved fluorescence from the oriented rods in the mixed films. However we did not observe any difference as compared with un-oriented samples. The time-resolved relaxation dynamics did not show any significant difference between the rod luminescence lifetime measured for polarization directions perpendicular and parallel to the rubbing direction. 6.4 Conclusions We have synthesized rod-shaped CdSe nanocrystals covered with a thin CdS shell and an aspect ratio of 3.6. The rods were almost monodisperse and showed a fluorescence

94 Chapter 6 in the red part of the spectrum with a high quantum yield (ηF = 45%). When mixed with green-emitting spherical CdSe/CdS nanocrystals in thin solid films, transfer of excited state energy from the spherical dots to the elongated rods occurs after photoexcitation. As a consequence, the fluorescence intensity and lifetime of the green-emitting dots decreases in favor of an increased red emission and lifetime. The Förster radius for energy transfer was determined to have an upper limit of 6.6 nm, close to the sum of the diameters of dots and rods. Hence, only dots that are in the direct vicinity of the rods can efficiently participate in energy transfer. The experimental result that the green fluorescence of the dots is quenched for 50% at a rod:dot ration of 1:30 is in good agreement with the geometrical constraints that 30-40 individual dots can assemble around one rod at the Förster distance. By simple rubbing it is possible to uniaxially align the rod shaped crystals in a mixed film with quantum dots. In these mixtures, the uniaxially aligned rods show polarized fluorescence. However as opposed to the behavior observed for films of pure rods, the direction of polarization of the light used for excitation of the mixed film had almost no effect on the emitted intensity. This indicates that excitation light was to a large extent absorbed isotropically by the dots which then transferred the excited state energy to the rods, which showed polarized light emission.

References 1 Alivisatos, A. P. Science 1996, 271,933 2 Bawendi, M. G.; Carroll, P. J.; Wilson, W. L.; Brus, L. E. J. Chem. Phys. 1992, 96,

946 3 Murray C. B.; Kagan, C. R.; Bawendi, M. G. Annu. Rev. Mater. Sci. 2000, 30, 545 4 Kim F.; Song, J.; Yang, P. J. Am. Chem. Soc. 2002, 124, 14316 5 Bruchez, M.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. Science 1998, 281,

2013 6 Chan, W. C. W.; Nie, S. Science 1998, 281, 2016 7 Wolcott, A.; Gerion, D.; Visconte, M.; Sun, J.; Schwartzberg, A.; Chen, S.; Zhang,

J. Z. J. Phys. Chem. B 2006, 110, 5779 8 Mulder, W. J. M.; Koole, R.; Brandwijk, R. J.; Storm, G.; Chin, P. T. K.; Strijkers,

G. J.; de Mello Donegá, C.; Nicolay, K.; Griffioen, A. W. Nano Lett. 2006, 6, 1 9 Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994, 370, 354 10 Dabboussi, B. O.; Bawendi, M. G.; Onitsuka, O.; Rubner, M. F. Appl. Phys. Lett.

1995, 66, 1316 11 Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002, 292, 2425

Energy Transfer and Polarized Emission in CdSe NC Solids 95

12 Peng, X.; Manna, L.; Yang, W.; Wickham, J.; Scher, E.; Kadavanich, A.;

Alivisatos, A. P. Nature 2000, 404, 59 13 Peng, Z. A.; Peng, X. J. Am. Chem. Soc. 2000, 123, 1389 14 Li, L. S.; Hu, J.; Yang, W.; Alivisatos, A. P. Nano Lett. 2001, 1, 349 15 Bunge, S. D.; Krueger, K. M.; Boyle, T. J.; Rodriquez, M. A.; Headley, T. J.;

Colvin, V. L. J. Mater. Chem. 2003, 13, 1705 16 Hu, J. T.; Li, L. S.; Yang, W. D.; Manna, L.; Wang, L. W.; Alivisatos, A. P.

Science 2001, 292, 2060 17 Talapin, D. V.; Koeppe, R.; Götzinger, S.; Kornowski, A.; Lupton, J. M.; Rogach,

A. L.; Benson, O.; Feldmann, J.; Weller, H. Nano Lett. 2003, 3, 1677 18 X. Chen, A. Nazzal, D. Goorsky, M. Xiao, Z. A. Peng, X. G. Peng, Phys. Rev. B.

2001, 64, 245304 19 Kan, S. H.; Mokari, T.; Rotheberg, T.; Banning, U. Nature Mater. 2003, 2, 155. 20 Shabaev, A.; Efros, A. L. Nano Lett. 2004, 4, 1821 21 Hikmet, R. A. M.; Chin, P. K. T.; Talapin, D. V.; Weller, H. Adv. Mater. 2005, 17,

1436 22 Montali, A.; Bastiaansen, C.; Smith, P.; Weder , C. Nature 1998, 392, 261 23 Kagen, C. R.; Murray, C. B.; Bawendi, M. G. Phys. Rev. B 1996, 54, 8633 24 Kagen, C. R.; Murray, C. B.; Nirmal, M.; Bawendi, M. G. Phys. Rev. Lett. 1996,

76, 1517 25 Wuister, S. F.; Koole, R.; de Mello Donegá, C.; Meijerink, A. J. Phys. Chem. B,

2005, 109, 5504 26 Wargnier, R.; Baranov, A. V.; Maslov, V. G.; Stsiapura, V.; Artemyev, M.; Pluot,

M.; Sukhanova, A.; Nabiev, I. Nano Lett. 2004, 4, 451 27 Anni, M.; Manna, L.; Cingolani, R.; Valerini, D.; Lomascolo, M. Appl. Phys. Lett.

2004, 85, 4169 28 Becker, K.; Lupton, J. M.; Müller, J.; Rogach, A. L.; Talapin, D. V.; Weller, H.;

Feldmann, J. Nat. Mater. 2006, 5, 777-781 29 Mokari, T.; Banin, U. Chem. Mater. 2003, 15, 3955 30 Dabbousi, B. O.; Rodriguez-Viejo, J.; Mikulec, F. V. ; Heine, J. R.; Mattoussi, H.;

Ober, R.; Jensen, K. F.; Bawendi, M. G. J. Phys. Chem. B 1997, 101, 9463 31 Talapin, D. V.; Koeppe, R.; Götzinger, S.; Kornowski, A.; Lupton, J. M.; Rogach,

A.; Benson, L. O.; Feldmann, J.; Weller, H. Nano Lett. 2003, 3, 1677 32 Manna, L.; Scher, E. C.; Li, L. S.; Alivisatos, A. P. J. Am. Chem. Soc. 2002, 124,

7136. 33 Meskers, S. C. J.; Peeters, E.; Langeveld-Voss, B. M. W.; Janssen, R. A. J. Adv.

Mater. 2000, 12, 589

96 Chapter 6 34 Strickler, S. J.; Berg, R. A. J. Chem. Phys. 1962, 37, 814 35 Karstens, T.; Kobs, K. J. Phys. Chem. 1980, 84, 1871 36 Eaton, D. F. J. Photochem. Photobiol. B 1988, 2, 523 37 Mekis, I.; Talapin, D. V.; Kornowski, A.; Haase, M.; Weller, H. J. Phys. Chem. B

2003, 107, 7454 38 Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993, 115, 8706 39 Yu, W. W.; Qu, L.; Guo, W.; Peng, X. Chem. Mater. 2003, 15, 2854 40 Valeur, B. Molecular .Fluorescence – An introduction: Principles and

Applications. VCH, Weinheim, 2000 41 Scholes, G. D.; Andrews, D. L. Phys Rev. B. 2005, 72, 125331 42 Oh, M. H. J.; Gentleman, D. J.; Scholes, G. D. Phys Chem. Chem. Phys. 2006, 8,

5079

97

Chapter 7

Highly Luminescent Ultra Thin Mn Doped ZnSe Nanowires

Summary Highly luminescent colloidal thin ZnSe nanowires with Mn dopants have been prepared using preformed Li4[Zn10Se4(SPh)16] or Li2[Zn4(SPh)10] clusters together with elemental selenium and manganese stearate at moderate temperatures. The reaction was tunable between spherical particles and anisotropic nanowire formation by changing the selenium content. The wire diameter could be changed between 1 and 3 nm, resulting in aspect ratios above 80 for 2.5 nm wide nanowires. The nanowires are highly crystalline and show a manganese photoluminescence. The emissive properties were further improved by the formation of a CdSe shell on the crystal surface, leading to colloidal nanowires with a luminescence quantum yield up to 40%. Flow-aligned Mn doped ZnSe nanowires showed a small polarized Mn emission with polarization perpendicular to the long axis of the nanowires.

98 Chapter 7 7.1 Introduction Colloidal semiconductor nanocrystals (NCs) with dimensions below the bulk exciton radius exhibit size related optical and electronic properties.1-3 Their stable, tunable, bright, and narrow photoemission make them favorable for application as emitters in biomedical labeling,4 LEDs,5,6 and lasers.7,8 By introducing transition metal dopants into the semiconductor NCs,9-11 their functionality can be further enhanced by combining quantum size effects and atomic band transitions, resulting in a novel class of materials. ZnSe has shown to be a very suitable host material for creating Mn doped NCs with various doping levels.10,12,13 Up to now, mainly spherical doped ZnSe colloidal NCs have been reported,11,13-15 while anisotropic Mn doped ZnSe materials have been produced by templated growth and thermal deposition techniques.16-18 In fact, only limited attention has been given to shape control of colloidal ZnSe nanocrystals. One dimensional (1D) ZnSe NCs such as nanowires19,20 and nanorods21 have been described, but these structures often have dimensions larger than the spatial confinement regime.22,23 Recently Panda et al.,24 demonstrated the formation of narrow nanorods and nanowires in a lyothermal synthesis at moderate temperatures.

Here we report a reproducible and stable solution based synthesis to create the first colloidal wurtzite ZnSe nanowires, with Mn doping. These colloidal ZnSe Mn doped nanowires have optical properties that depend on size, shape, and doping level in one single crystal. The ZnSe:Mn nanowires were prepared by using manganese stearate and pre-formed Li4[Zn10Se4(SPh)16] single source clusters or from Li2[Zn4(SPh)10] clusters and elemental selenium.25,26 The latter method enabled the control over the zinc:selenium ratio in the synthesis. The ZnSe:Mn nanowires are highly crystalline and can be synthesized with ultra-narrow diameters from 1 to 3 nm, well below the bulk exciton Bohr radius (3.8 nm) of ZnSe. The luminescent properties of these nanowires were further improved by the formation of CdSe on the crystal surface increasing the photoluminescence quantum yield up to 40%. These doped nanowires combine quantum size and shape effects with atomic band transitions, making them particular interesting for anisotropic optical studies and for anisotropic spin related studies.

7.2 Experimental Chemicals. Hexadecylamine (HDA, 90%), trioctylphosphine (TOP), trioctylphosphine oxide (TOPO), tributylphosphine (TBP), and anhydrous solvents (toluene, methanol, chloroform) were all purchased from Aldrich. Selenium powder (99.99%, 200 mesh) and

Highly Luminescent Ultra Thin Mn Doped ZnSe Nanowires 99

anhydrous cadmium acetate (99.99+%) were purchased from Chempur. LiNO3, Zn(NO3)2·6H2O, thiophenol, and triethylamine were purchased from Fluka. All reagents were used as received except TOPO and HDA. These reagents were dried and degassed before use by heating under vacuum at 393 K for 1.5 h. Cluster synthesis. Li2[Zn4(SPh)10] and Li4[Zn10Se4(SPh)16] clusters were synthesized using a modification of the method proposed by Dance et al.25 First, Zn(NO3)2·6H2O (10.5 g, 35 mmol) dissolved in methanol (35 mL) was added to a well-stirred mixture of thiophenol (9.3 g, 91 mmol) and triethylamine (12.7 mL, 91 mmol) in methanol (20 mL) at room temperature. Subsequently LiNO3 (2.8 g, 40 mmol) in methanol (20 mL) was added and stirring was continued until all precipitate had dissolved. The Li2[Zn4(SPh)10] clusters were crystallized by cooling to 255 K and isolated by filtration and drying under vacuum for 24 h. The selenium containing cluster Li4[Zn10Se4(SPh)16] was formed using elemental selenium powder (1.64 g, 20.8 mmol) which was added to a solution of Li2[Zn4(SPh)10] (17 g, 13 mmol) in DMF (40 mL) and allowed to stir for 4 h at room temperature. The solution was filtered and methanol was added until a precipitate was observed, subsequently the mixture was stored a 253 K for 12 h for further crystallization. The crystals were filtered and washed with cold acetonitrile (273 K) and dried under vacuum. ZnSe:Mn synthesis using Li4[Zn10Se4(SPh)16]. Highly crystalline spherical and anisotropic ZnSe and Mn doped ZnSe NCs were prepared following a modification of the method published by Strouse et al.26 HDA (2.5 g) and manganese stearate (5.7 mg, 9.2 μmol for 5 % doping to Zn per mol) were mixed and dried for 1.5 h under vacuum at 383 K. After cooling to room temperature, Li4[Zn10Se4(SPh)16] (50 mg, 18.22 μmol) was added, followed by slow heating under vacuum to 353 K. At 353 K the mixture was flushed with argon and then heated slowly (4 K/min) to 563 K. At 563 K the average particle diameter was about 2.5 nm. Aliquots of the growth solution were taken at regular time intervals during the shell growth and cooled into room temperature chloroform. No post-preparative procedures were employed after sampling, except for the samples to be investigated by transmission electron microscopy, which were purified by precipitation with anhydrous methanol. ZnSe:Mn synthesis using Li2[Zn4(SPh)10] and Se. In a typical synthesis for Mn doped particles, HDA (2.5 g), selenium powder (2.9 mg, 36.7 μmol), and manganese stearate (5.7 mg, 9.2 μmol for 6 % doping to Zn per mol) were mixed and dried for 1.5 h under vacuum at 383 K. After cooling to room temperature, Li2[Zn4(SPh)10] (49 mg, 35.9 μmol) was added, followed by slow heating under vacuum to 353 K. All following steps were identical to the synthesis using Li4[Zn10Se4(SPh)16] described above. ZnSe:Mn synthesis with variable Se concentration. The synthesis using the Li2[Zn4(SPh)10] clusters enabled the variation of the amount selenium in the reaction

100 Chapter 7 mixture. To study the effect of the amount of selenium on the formation of either wires or spherical particles, different amounts of elemental selenium were combined with Li2[Zn4(SPh)10], all other reaction conditions being identical as described above. Surface passivation with CdSe. The ZnSe or ZnSe:Mn NCs were first purified by precipitation with anhydrous methanol from a chloroform solution. After centrifugation, the purified NCs ( ≈21 mg ZnSe:Mn) were dissolved in toluene (1 mL) and redispersed in a TOPO (67 wt %; 4 g) / HDA (33 wt %; 2 g) mixture at 333 K under vacuum. The surface passivation by CdSe was carried out in the temperature range 423 to 493 K, by dropwise (1 drop every 3 min) addition of 200 μL of a solution of cadmium acetate in TOP (0.12 mol/L) and 200 μL of Se in TBP (0.12 mol/L). Optical spectroscopy. Absorption and photoluminescence (PL) spectra were recorded with a Perkin–Elmer Lambda 900 spectrophotometer and an Edinburgh Instruments FS920 spectrophotometer, respectively. PL measurements were carried out on samples with a low optical density (≤0.1 at the excitation wavelength). PL quantum yields were determined using rhodamine 101 in 0.01% HCl ethanolic solution as reference, following the method reported in ref. 27 Transmission electron microscopy. TEM and high-resolution TEM were performed on a TECNAI G2 20 transmission electron microscope (FEI Co., The Netherlands) operated at 200 kV The samples for TEM were prepared by placing a drop of chloroform solution of NCs on a carbon coated copper (400-mesh) TEM grid. The excess liquid was removed by blotting using filter paper. X-ray diffraction. XRD was recorded with a Rigaku D/Max B diffractometer (graphite monochromator, scintillation counter) using Cu Kα radiation (1.54056 Å) operating at 40 kV, 30 mA. The data were collected from drop cast NCs on glass samples using a step scan mode with a step size of 0.02º and a counting time of 3 s per step in the range 2θ between 20° to 80º Elemental analysis ICP-AES. An amount of the crude reaction mixture containing ZnSe:Mn nanowires (~20 mg) was dissolved in anhydrous chloroform (2 mL) and purified by precipitation with anhydrous methanol, followed by redispersion in anhydrous chloroform (1 mL) and precipitation with anhydrous methanol. After removing the remaining solvents by evaporation under argon was a 63% HNO3 solution (45 μl) was used to the dissolve the ZnSe:Mn nanowires. The solution was further diluted to 25.0 mL with pure water. The ratio between Zn and Mn in the samples was measured using an ICP-AES spectrometer against a reference line. Electron paramagnetic resonance spectroscopy. EPR spectra were recorded with a Bruker ESP 300E spectrometer operating at X-band (9.43 GHz) . The sample was held at 120 K in a quartz tube.

Highly Luminescent Ultra Thin Mn Doped ZnSe Nanowires 101

7.3 Results and Discussion

ZnSe:Mn nanoparticles were prepared by slow heating of a mixture, of Li4[Zn10Se4(SPh)16] and manganese stearate (5 mol% Mn compared to Zn) in HDA from 353 K to 563 K during 1 h. The particle growth is clearly reflected in the changes in the absorption and photoluminescence (PL) spectra (Figure 1) that exhibit a red shift with increasing temperature, corresponding to an increase in particle diameter. The nanocrystals isolated at 543 K show manganese emission at 590 nm with PL quantum yield of about 3%. This emission of ZnSe:Mn is assigned to an internal Mn2+ transition (4T1 → 6A1).13 At ~380 nm a weak ZnSe exciton emission is visible (Figure 1b) that has a small Stokes shift compared to the corresponding exciton absorption (Figure 1a).

During the initial stages of the particle growth, the absorption spectra show no sharp exciton absorption peak. Instead, an absorption peak between 310 and 320 nm is present, well below the onset of the absorption at ~340 nm. This peak is expected to originate from very small, so called magic sized, ZnSe clusters with a preferred thermodynamic stability28,29 that eventually disappear with increasing reaction time and temperature.

The PL intensity of the exciton and manganese emission show a significant increase upon particle growth (Figure 1b). The position of the maximum of the manganese emission varies during the different stages of particle growth (Figure 2a). Initially there is a small blue shift from 610 nm (sample 443 K) to 590 nm (sample 523 K). It has been established that Mn2+ ions in a ZnSe host matrix show a typical luminescence maximum at 590 nm,12 which shifts to 610 nm in the presence of thiolate ligands,30 as a result of a change in crystal field splitting. Hence the experiment in Figure 2a may indicate that the Mn2+ ions become effectively shielded from the surface, which is capped by thiophenols,31 upon growth of the ZnSe particle. The emission of particles heated above 543 K shows again a small red shift which can be caused by decomposition of the thiophenol surfactants that decompose at high temperatures,32 introducing sulfide ions in ZnSe host crystal, leading to a change in crystal field. The efficiency of Mn doping was confirmed by ICP-AES elemental analysis, which showed that the doping level in ZnSe:Mn NCs corresponds to the ratios of Li2[Zn4(SPh)10] and manganese stearate added.

102 Chapter 7

300 320 340 360 380 4000.0

0.1

0.2

0.3

Nor

mai

zed

Abso

rban

ce

Wavelength (nm)

443 K 463 K 483 K 503 K 523 K 543 K 563 K

(a)

350 400 450 500 550 600 6500123456789

1011

PL in

tens

ity (a

.u.)

Wavelength (nm)

483 K 503 K 523 K 543 K 563 K

(b)

Figure 1. Evolution of the absorption (a) and PL (b) spectra during the growth of ZnSe:Mn particles with increasing temperature.

To obtain additional evidence that the Mn2+ ions are actually embedded in the

ZnSe host matrix, we studied the nanocrystals with EPR spectroscopy. The X-band EPR spectrum at 120 K shows a six line spectrum superposed on a broad background (Figure 2b). The six line spectrum results from the hyperfine interaction of the unpaired electrons with the 55Mn nuclear spin (I = 5/2). The 55Mn hyperfine splitting is sensitive to the ions surrounding the Mn center, and can be used to determine the position of the Mn atoms in the crystal. The hyperfine splitting of 63.0×10-4 cm-1 determined from the EPR spectrum is in close agreement with literature data for Mn in bulk ZnSe (61.7×10-4 cm-1).33,34 When the Mn2+ ions are located on the particle surface, an increase of the hyperfine splitting to 76.6×10-4 cm-1 is expected.14,15 The broad background in Figure 2b likely results from the relative high Mn concentration ≈1.4%, which results in dipole-dipole interactions between the Mn centers that broaden the spectra.34

Highly Luminescent Ultra Thin Mn Doped ZnSe Nanowires 103

The EPR measurement and the blue shifted Mn emission at 590 nm lead to the

conclusion that at reaction temperatures between 523 and 543 K most of the Mn atoms are located inside the ZnSe host.

500 550 600 650 7000

2

4

6

8

10

Nor

mal

ized

PL

inte

nsity

Wavelength (nm)

443 K

563 K

523 K

(a)

3000 3100 3200 3300 3400 3500 3600 3700

EP

R In

tens

ity (a

.u.)

Magnetic field (G)

(b)

Figure 2. (a) Normalized PL spectra of the Mn emission taken at different stages during particle growth. (b) EPR spectrum for ZnSe:Mn nanowires.

The X-ray diffraction pattern of the ZnSe:Mn NCs is shown in Figure 3a. Within

the angular resolution of the experiment, wurtzite and zinc blende structured ZnSe have several overlapping diffraction peaks, especially the (002), (110), and (112) peaks of wurtzite are close to the (111), (220), and (311) peaks of zinc blende. However the (100)

104 Chapter 7 and (101) peaks for wurtzite in Figure 3a, have no equivalent for zinc blende and give credence to assume that the ZnSe:Mn NCs have a wurtzite structure. On the other hand, some weak peaks expected for wurtzite, are not visible, viz. the (102) at ~36 deg and the (103) at ~50 deg. This may be caused by attenuation of the (102) and (103) peaks by the combination of a small diameter, resulting also in broad (110) and (112) peaks, and some stacking faults producing short zinc blende domains. The high intensity of the (002) peak together with broad (110) and (112) peaks, suggest an anisotropic growth in the (002) direction with a narrow particle width.

The shape and dimensions of the particles were further investigated with TEM (Figure 3b). This revealed that a mixture of spherical and rod shaped particles was formed using the pre-formed Li4[Zn10Se4(SPh)16] clusters, the width of the rods and the diameter of the spherical particles was the same, explaining the clear exciton absorption peak in the absorption and spectrum (Figure 1a). The nanowire diameter observed by TEM increased from ~1 to ~3 nm (563 K, 1 h) with increasing reaction time and temperature.

20 30 40 50 600

100

200

300

400

500

Inte

nsity

(a.u

.)

2θ

100

002

101110

103 112

(a)

50 nm50 nm 50 nm50 nm Figure 3. (a) Powder X-ray (λ =1.54056 Å) diffraction pattern of ZnSe:Mn (5% Mn:Zn). (b) TEM image of Mn doped ZnSe nanodots and nanorods (scale bar is 50 nm). (c) TEM image of ZnSe:Mn nanorods made with extra TBP-Se (scale bar is 50 nm), the inset shows a high resolution TEM image (scale bar is 10 nm).

10 nm10 nm

b c

Highly Luminescent Ultra Thin Mn Doped ZnSe Nanowires 105

In a second experiment, the concentration of selenium in the reaction mixture was

increased by adding tributylphosphine selenium (TBPSe) to the Li4[Zn10Se4(SPh)16] clusters at the start of the reaction to obtain a 1:1 Zn:Se molar ratio. This resulted in the formation of mainly ZnSe:Mn nanorods with aspect ratios up to 17 (Figure 3c). The high resolution images show that the rods are single crystalline.

As we wished to study the influence of the selenium concentration on the formation of ZnSe and ZnSe:Mn doped spherical or rod-shaped nanocrystals in more detail but found that TBPSe was not very reactive, we used Li2[Zn4(SPh)10] clusters containing only Zn, instead of pre-formed Li4[Zn10Se4(SPh)16] clusters having both Zn and Se, and combined this with elemental selenium and manganese stearate (5 mol% Mn compared to Zn) in a HDA solution. Heating this mixture from room temperature resulted in a rapid incorporation of selenium into the clusters. The particle growth and formation took place during the following hour, when the reaction mixture was heated to 563 K (4 K/5 min).

Figure 4. TEM images of ZnSe:Mn nanocrystals synthesized with different Zn:Se ratio.

Figure 4 shows TEM images of the particles formed with different Zn:Se ratios. At

high Zn:Se ratios (1:0.1), predominantly spherical particles were formed, but formation of nanowires rapidly set in when the amount of Se was increased. The highest yield of nanowires was obtained for a Zn:Se ratio of 1:0.3. Together with the increase in wire formation also the length and aspect ratio of the wires increased to 200 nm and 80, respectively. When the selenium concentration was further increased, spherical particles started to appear again and at a Zn:Se ratio of 1:1 only spherical particles were formed. At

106 Chapter 7 even higher Se concentration, uncontrolled anisotropic growth occurred. We note that the manganese concentration showed no influence on the formation of the NCs; without Mn similar ratios between spherical and rod shaped crystals were found.

The formation of the nanowires as shown in Figure 4 may involve different mechanisms, i.e. oriented attachment of small crystals35-37 or specific binding of surfactants to different facets of wurtzite crystals, giving rise to oriented crystal growth.24 The formation of nanowires by oriented attachment can occur when dipoles are created in the ZnSe cluster when opposite crystal facets are terminated with zinc or selenium. This mechanism could be operative here, because the dot versus wire formation strongly depends on the Zn:Se ratio. As the reactivities of the zinc and selenium source are not the same, an optimum nanowire formation is found at: 1:0.3 Zn:Se. A further indication for this mechanism is the fact that at all Zn:Se ratios used there is always a certain fraction of spherical particles formed. This is consistent with the notion that oriented attachment is a statistical effect and not all ZnSe clusters will form wires. The other mechanism, anisotropic growth,38,39 has been previously been suggested by Panda et al.24 for wurtzite ZnSe nanowires made from zinc acetate and selenourea in the presence of weak binding octadecylamine. Here anisotropic growth was explained by the stronger binding of the octadecylamine to the long side walls and the zinc terminated tips of the rods, compared to the selenide terminated tip, forcing one dimensional growth. This growth mechanism implies that the kinetics of the alkylamine adsorption to the mineral surface are significantly slower than the growth of the ZnSe mineral phase.24 Hence, also here the Zn:Se concentration and reactivity will be important parameters.

To increase the PL quantum yield of the ZnSe:Mn NCs, the surface was passivated by the formation of CdSe on the NC surface. The CdSe shell growth was performed in a mixture of HDA and TOPO at 493 K, using anhydrous cadmium acetate dissolved in TOP and TBPSe. The alternate addition of Se and Cd resulted immediately in a strong increase luminescence after the first addition. Successive alternate additions of Se and Cd resulted in a further increase of PL quantum yield up to ~40%. (~24 μmol Se and ~25 μmol Cd to ~145 μmol (21 mg) of ZnSe:Mn NCs). The effect of addition either Se or Cd was studied in more detail by monitoring the absorption and luminescence spectra at time intervals of ~15 min. between each addition at 493 K. The PL spectrum (Figure 5a) reveals that the major increase in luminescence intensity occurs upon the addition of Cd whereas the addition of Se only results in a much smaller increase in PL intensity. The PL increase by adding Cd, however, is somewhat variable as can be seen from the small decrease when the second aliquot of Cd is added. The third aliquot causes a twofold increase, resulting in a total 55 fold increase in Mn emission intensity compared to the original ZnSe:Mn core emission. The absorption spectrum shows also a clear red shift with increasing CdSe shell

Highly Luminescent Ultra Thin Mn Doped ZnSe Nanowires 107

growth and, again the major red shift is observed when Cd is added, while addition of Se has no significant influence on the absorption spectrum. (Figure 5b)

300 400 500 600 700 8000

1

2

3

4

5

6

7

8

f. 104 μmol Cd

b. 34 μmol Cd

c. 53 μmol Se

PL in

tens

ity (c

ount

s/10

4 )

Wavelength (nm)

a. 17μmol Se

e. 85 μmol Sed. 69μmol Cd

ZnSe:Mn core

(a)

300 350 400 4500.00

0.05

0.10

0.15

0.20

0.25

0.30

Nor

mal

ized

Abs

orba

nce

Wavelength (nm)

ZnSe:Mn core a Se b Cd c Se d Cd e Se f Cd

(b)

Figure 5. PL (a) and absorption (b) spectra from ZnSe:Mn:CdSe nanowires at different stages of CdSe shell growth. The letters a to f correspond to the successive order of additions of either CdAc or TOP-Se, the amounts of Cd or Se are the total amounts added to the reaction.

After passivating the nanowires with a thin layer of CdSe the PL quantum yield

increased from ~ 3% to 40%. After size selective precipitation to remove the smaller quantum dots, the nanowires remained highly luminescent and showed no remarkable difference in quantum yield.

108 Chapter 7 The absorption and luminescence of anisotropic NCs is known to show a clear polarization.40-42 To study anisotropic optical properties of dopant emission, we aligned size selected ZnSe:Mn:CdSe nanowires using a 1 mm flow cell. Alignment of the 1D wires along the flow direction in the cell was achieved by applying a continuous flow of ZnSe:Mn:CdSe nanowires dispersed in cyclohexane. The PL spectra obtained for the excitonic and atomic Mn emission were corrected for both the polarization dependence of light source and detector caused by the gratings. The corrected polarized PL spectra (Figure 6) show a weak polarization dependence of the Mn emission. The highest PL intensity is found perpendicular to the flow direction, and hence orthogonal to the long axis of the nanowires.

400 450 500 550 600 650 7000.0

0.5

1.0

1.5

2.0

2.5

PL in

tens

ity (c

ount

s/10

3 )

Wavelength (nm)

|

//

Figure 6. Polarized PL spectra from macroscopic aligned nanowires in cyclohexane, measured parallel and perpendicular to the flow direction. The spectra were recorded with a polarized excitation source at 366 nm.

A polarization dependence of both photoexcitation and photoemission has previous

been shown for CdSe, CdS, and InAs nanorods and nanowires.6,40,43-45 The emission in these materials however is polarized along the long axis of the rod or wire. In contrast the Mn emission in ZnSe nanowires is polarized perpendicular to the long axis. This is consistent with the results obtained by Golan et al.,42 where the highest polarization dependence in narrow wurtzite ZnSe wires was found perpendicular to the long axis.

Highly Luminescent Ultra Thin Mn Doped ZnSe Nanowires 109

7.4 Conclusion

We synthesized narrow manganese doped ZnSe nanowires with an aspect ratio above 80. The wire formation is controlled by the ratio between Zn and Se in the reaction mixture. The ZnSe:Mn wires are highly monodisperse in the width direction and show an almost pure Mn emission at 590 nm. PL and EPR spectroscopy indicate that the manganese dopant is incorporated in the ZnSe matrix, rather than at the surface. The PL quantum yield of the ZnSe:Mn nanowires is 40% after passivation with a thin CdSe shell. Macroscopic alignment of the nanowires in a continuous flow cell revealed a small polarization of the manganese emission perpendicular to the long axis. The nanowires presented in this work reveal the combined effects of size, shape, shell, and transition metal doping on the photoluminescence in a single nanocrystal.

References

1 Alivisatos, A. P. J. Phys. Chem. 1996, 100, 13226 2 Alivisatos, A. P. Science 1996, 271, 933 3 Brus, L. E. J. Chem. Phys. 1983, 79, 5566 4 Michalet, X.; Pinaud, F. F.; Bentolila, L. A.; Tsay, J. M.; Doose, S.; Li, J. J.;

Sundaresan, G.; Wu, A. M.; Gambhir, S. S.; Weiss, S. Science 2005, 307, 538- 5 Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994, 370, 354 6 Hikmet, R. A. M.; Chin, P. T. K.; Talapin, D. V.; Weller, H. Adv. Mater. 2005,

17, 1436 7 Klimov, V. I.; Mikhailovski, A. A.; Xu, S.; Malko, A.; Hollingsworth, J. A.;

Leaterdale, C. A Eisler, H. J.; Bawendi, M. G. Science 2000, 290, 314 8 Kazes, M.; Lewis, D. Y.; Ebenstein, Y.; Mokari, T. Banin, U. Adv. Mater. 2002,

14, 317 9 Becker, W. G.; Bard, A. J. J. Phys. Chem. 1983, 87, 4888 10 Bhargava, R. N.; Gallagher, D. ; Hong, X.; Nurmikko, A. Phys. Rev. Lett. 1994,

72, 416 11 Pradhan, N.; Goorskey, D.; Thessing, J.; Peng, X. J. Am. Chem. Soc. 2005, 127,

17586 12 Suyver, J. F.; Wuister, S. F.; Kelly, J. J.; Meijerink A. Phys. Chem. Chem. Phys.

2000, 2, 5445 13 Norris, D. J.; Yao, N.; Charnock, F. T.; Kennedy, T. A. Nano Lett. 2001, 1, 3 14 Hofmann, A.; Graf, C.; Boeglin, C.; Rühl, E. ChemPhysChem 2007, 8, 2008 15 Norris, D. J.; Efros, A. L.; Erwin, S. C. Science 2008, 319, 1776 16 Ramanathan, S.; Patibandla, S.; Bandyopadhyay, S.; Anderson, J.; Edwards, J. D.

Nanotechnology 2008, 19, 195601 (6pp). 17 Radovanovic, P. V.; Barrelet, C. J.; Gradečak, S.; Qian, F.; Lieber C. M. Nano

Lett. 2005, 5, 1407

110 Chapter 7

18 Jia, T. Baba, M.; Huang, M.; Zhao, F.; Qiu, J.; Wu, X.; Ichihara, M.; Suzuki, M.; Li, R.; Xu, Z.; Kuroda, H. Solid State Commun. 2007, 141, 635

19 Duan, X. F.; Lieber, C. M. Adv. Mater. 2000, 12, 298 20 Xiang, B.; Zhang, H. Z.; Li, G. H.; Su, F. H.; Wang, R. M.; Xu, J.; Lu, G. W.;

Sun, X. C.; Zhao, Q.; Yu, D. P. Appl. Phys. Lett. 2003, 82, 3330 21 Cozzoli, P. D.; Manna, L.; Curri, L.; Kudera, S.; Giannini, C.; Striccoli, M.;

Agostiano, A Chem. Mater. 2005, 17, 1296 22 Lv, R.; Cao, C.; Zhai, H.; Wang, D.; Liu, S.; Zhu, H.; Solid State Commun. 2004,

130, 241 23 Zhu, Y. C.; Bando, Y. Chem. Phys. Lett. 2003, 377, 367 24 Panda, A. B.; Acharya, S.; Efrima, S. Adv. Mater. 2005, 17, 2471 25 Dance, I. G.; Choy, A.; Scudder, M. L. J. Am. Chem. Soc. 1984, 106, 6285 26 Cumberland, S. L.; Hanif, K. M.; Javier, A.; Khitrov, G. A.; Strouse, G. F.;

Woessner, S. M.; Yun, C. S. Chem. Mater. 2002, 14, 1576 27 de Mello Donegá, C.; Hickey, S. G.; Wuister, S. F.; Vanmaekelbergh, D.;

Meijerink, A. J. Phys. Chem.B 2003, 107, 489

28 Rogach, A. L.; Katsikas, L; Kornowski, A; Su, D; Eychmüller, A; Weller, H. Ber. Bunsen Ges. Phys. Chem.1997, 101, 1668

29 Wuister, S. F.; Driel, F.; Meijerink, A. J. Luminescence 2003, 102, 327 30 Pradhan, N.; Peng, X. J. Am. Chem. Soc. 2007, 129, 3339 31 Chin, P. T. K.; Stouwdam, J. W. ; van Bavel, S. S.; Janssen, R. A. J.

Nanotechnology 2008, 19, 205602 (1-8) 32 Archer, P. I.; Santangelo, S. A.; Gamelin, D. R. J. Am. Chem. Soc. 2007, 129,

9808 33 Ludwig, G. W.; Woodbury, H. H. Electron Spin Resonance in Semiconductors.

In Solid State Physics; Seitz, F., Turnbull, D., Eds.; Academic Press: New York, 1962, 13, 298.

34 Kennedy, T. A.; Glaser, E. R.; Klein, P. B.; Bhargava, R. N. Phys. Rev. B 1995, 52, R14356-

35 Cho, K.-S.; Talapin, D. V.; Gaschler. W.; Murray C. B. J. Am. Chem. Soc. 2005, 127, 7140

36 Pradhan, N.; Xu, H.; Peng, X. Nano Lett. 2006, 6, 720 37 Ribeiro, C.; Lee, E. J. H.; Longo, E.; Leite, E. R. ChemPhysChem. 2006, 7, 664 38 Yu, W. W.; Wang, Y. A.; Peng, X. G. Chem. Mater. 2003, 15, 4300 39 Manna, L.; Wang, L. W.; Cingolani, R.; Alivisatos, A. P. J. Phys. Chem. B 2005,

109, 6183 40 Hu, J. T.; Li, L. S.; Yang, W. D.; Manna, L.; Wang, L. W.; Alivisatos, A. P.

Science, 2001, 292, 2060 41 Manna, L.; Scher, E. C.; Alivisatos, A. P. J. Clust. Sci. 2002, 13, 521 42 Somobrata, A.; Panda, A. B.; Efrima, S.; Golan, Y. Adv. Mater. 2007, 19, 1105 43 Peng, X.; Manna, L.; Yang, W.; Wickham, J.; Scher, E.; Kadavanich, A.;

Alivisatos, A. P. Nature 2000, 404, 59 44 Talapin, D. V.; Koeppe, R.; Götzinger, S.; Kornowski, A.; Lupton, J. M.;

Rogach, A. L.; Benson, O.; Feldmann, J.; Weller, H. Nano Lett. 2003, 3, 1677

Highly Luminescent Ultra Thin Mn Doped ZnSe Nanowires 111

45 Kan, S. H.; Mokari, T.; Rotheberg, T.; Bannin, U. Nature Mater. 2003, 2, 155

112 Chapter 7

113

Summary

This thesis focuses on various aspects of nanocrystals (NCs) based on transition metal

chalcogenide semiconductors (e.g. CdSe, CdTe, ZnSe). Different synthetic approaches have been developed for creating high quality monodisperse luminescent colloidal nanocrystals. The synthesis of these different nanosized crystals opened the opportunity to study and exploit the fascinating size-dependent physical and optical behavior of semiconductors in the nanosize regime. Detailed studies were performed on resonant energy transfer in host-guest systems between conjugated polymers and isotropic NCs, and between isotropic NCs and anisotropic NCs. The distinct optical properties make these luminescent NCs particularly interesting as the emissive component in lighting applications. These materials were therefore studied as the emissive component in thin film light-emitting diodes (LEDs). Some of the salient results are described in the following paragraphs.

The thesis starts with a general introduction on the history and general properties of colloidal semiconductor nanocrystals in chapter 1. The physical principles of size quantization are briefly explained and followed by an introduction on the effects of the particle shape and composition on the optical and electronic properties. The application of nanocrystals in luminescent devices is discussed, setting the stage for the overall aim of the thesis.

Chapter 2 describes a study of energy transfer in a host-guest system consisting of a blue-emitting poly(2,7-spirofluorene) (PSF) donor polymer and red-emitting CdSe/ZnS core-shell quantum dots as acceptor in solid films, using time-resolved optical spectroscopy, and in electroluminescent diodes. By introducing an electron transport layer in the LED the dominant pathway for quantum dot emission could be modified from energy transfer from the polymer host to direct electron-hole recombination on the quantum dot. This resulted in an increased device efficiency to 0.32 cd/A.

The preparation of highly luminescent, anisotropic CdTe/CdSe colloidal heteronanocrystals is described in chapter 3. The reaction conditions used (low temperature, slow precursor addition, and surfactant composition) resulted in a tunable shape from prolate to branched CdTe/CdSe nanocrystals. Upon CdSe shell growth, the heteronanocrystals show a gradual evolution from Type-I (direct recombination holes and electrons in one material) to Type-II (indirect recombination of holes and electrons at the interface of two materials) optical behavior. These heteronanocrystals show a remarkably high photoluminescence quantum yield (up to 82%) and negligible thermally induced

114 Summary quenching up to temperatures as high as 373 K. Such high quantum yields and stability are unprecedented for Type-II nanocrystals.

Chapter 4 shows a novel synthesis leading to highly luminescent CdTe nanocrystals using Li2[Cd4(SPh)10] clusters as a reactive Cd cluster compound at relatively low temperature, making it a safe precursor for the large scale synthesis of CdTe nanocrystals. The nanocrystals show high luminescent quantum yields up to 37% for branched CdTe nanostructures, and as high as 52% for CdTe/CdS core-shell heterostructures. CdTe/CdS nanocrystals were used to make LEDs in combination with organic layers for electron and hole injection. The devices show a maximum luminance efficiency of 0.35 cd/A.

The first NC LEDs that emit linearly polarized light using macroscopically oriented quantum rods are described in chapter 5. In these devices a thin layer of quantum rods with an aspect ratio of 2.5, which were macroscopically oriented by a simple rubbing technique, has been used as an emitter. Devices were constructed by sandwiching the oriented quantum rods between two organic layers with electron and hole conducting properties to obtain improved injection and emission properties. In this way a polarized LED with an emission at 620 nm, luminance efficiency of 0.65 cd/A, and external a quantum efficiency of 0.49% was obtained. The intensity of the electroluminescent light polarized in the direction parallel to the long axis of the rods was 1.5 times higher than in the perpendicular direction.

Excited state energy transfer from spherical green-emitting nanocrystals as donor to rod-shaped red-emitting nanocrystals as acceptor is demonstrated in chapter 6. For this purpose highly luminescent red-emitting core-shell CdSe/CdS quantum rods were synthesized and mixed with green-emitting core-shell CdSe/CdS quantum dots. For this donor-acceptor combination the Förster distance is less than 6.6 nm, which is close to sum of the diameters of the dots and rods. Hence, only quantum dots directly neighboring a quantum rod will efficiently participate in the energy transfer. A simple rubbing technique was used to uniaxially align the quantum rods dispersed in thin films of quantum dots. Such mixed films showed polarized red emission, while the excitation remained unpolarized.

Highly luminescent colloidal narrow ZnSe:Mn doped nanowires are prepared in chapter 7, using preformed Li4[Zn10Se4(SPh)16] and Li2[Zn4(SPh)10] clusters together with elemental selenium and manganese stearate at moderate temperatures. The nanowires are highly crystalline and show a bright manganese photoluminescence. The wire diameter could be changed between 1 and 3 nm, resulting in aspect ratios above 80 for 2.5 nm wide nanowires. The emissive properties were further improved by the formation of a CdSe shell on the ZnSe surface, leading to colloidal nanowires with a luminescence quantum yield up to 40%. The reaction was tunable between spherical particles and anisotropic nanowire formation by changing the selenium content. Aligned ZnSe:Mn doped

Summary 115 nanowires in a flow cell showed a weak polarized Mn emission with polarization perpendicular to the long axis of the nanowires.

In conclusion, the work described in this thesis show several syntheses of highly luminescent semiconductor NCs, with high control over size, shape, and composition. The exciting and distinct optical properties of these NCs have been studied, in combination with the energy transfer between the NCs and polymers. This enabled the creation of both nanophosphors and LEDs, exhibiting the same exciting optical properties but then on a macroscopic scale.

116 Summary

117

Samenvatting Het onderzoek beschreven in dit proefschrift was gericht op aspecten van overgangsmetaal-chalcogenide halfgeleider nanokristallen (bv. CdSe, CdTe, ZnSe). Diverse syntheseroutes voor de vorming van hoge kwaliteit, luminescente nanokristallen met een uniforme grootte zijn ontwikkeld. De kristallen met afmetingen van enkele nanometers vertonen fascinerende grootteafhankelijke fysische en optische eigenschappen, die in dit proefschrift bestudeerd en toegepast worden. De energieoverdracht in donor-acceptor systemen tussen geconjugeerde polymeren en bolvormige nanokristallen en tussen bolvormige en staafvormige nanokristallen is in detail onderzocht. De karakteristieke optische eigenschappen maken deze nanokristallen bijzonder interessant voor toepassing als actieve component in verlichting. De materialen zijn daarom onderzocht als emitter in dunne-film lichtgevende diodes (LEDs). De belangrijkste resultaten van het onderzoek worden hieronder beschreven. Het proefschrift begint met een algemene introductie over de geschiedenis en eigenschappen van colloïdale halfgeleider nanokristallen in hoofdstuk 1. De fysische achtergrond van kwantisering bij afnemende grootte wordt kort uitgelegd en gevolgd door een inleiding in de effecten van de deeltjesvorm en -samenstelling op de optische en elektronische eigenschappen. Het gebruik van nanokristallen voor verlichtingstoepassingen wordt beschreven en daarmee het uiteindelijke doel het van dit proefschrift. Hoofdstuk 2 beschrijft het onderzoek naar de energieoverdracht in de vaste stof tussen een blauwemitterend poly(2,7-spirofluoreen) (PSF) polymeer als donor en een roodemitterende CdSe/ZnS kern-schil kwantumstip als acceptor, gebruikmakend van tijdsopgeloste optische spectroscopie en toegepast in LEDs. Door het introduceren van een elektronentransportlaag in de LED wordt het overheersende mechanisme voor emissie van de kwantumstip gewijzigd. Zonder elektronentransportlaag overheerst de resonante energieoverdracht vanaf het polymeer, maar met elektronentransportlaag is directe recombinatie van elektronen en gaten op de kwantumstip het belangrijkste mechanisme. Dit resulteerde in een toegenomen efficiëntie van de LED tot 0,32 cd/A. De vorming van hoogluminescente, anisotrope CdTe/CdSe colloïdale hetero-nanokristallen wordt beschreven in hoofdstuk 3. De gebruikte reactiecondities (lage temperatuur, langzame toevoeging van de reactanten en surfactantsamenstelling)

118 Samenvatting resulteren in de overgang van langwerpige naar vertakte CdTe/CdSe nanokristallen. Met de toename van de dikte van de CdSe schil vindt een geleidelijke overgang plaats van Type-I (directe recombinatie van gaten en elektronen in één materiaal) naar Type-II (indirecte recombinatie van gaten en elektronen over het grensvlak tussen de twee materialen) optisch gedrag van de nanokristallen. Deze hetero-nanokristallen vertonen een opvallend hoge kwantumopbrengst voor luminescentie (>82%), gecombineerd met een verwaarloosbare temperatuurgeactiveerde doving voor temperaturen tot 373 K. Een dergelijke hoge kwantumopbrengst en stabiliteit zijn nog niet eerder gemeld voor Type-II kwantumstippen. Hoofdstuk 4 beschrijft een nieuwe syntheseroute om hoogluminescente CdTe nanokristallen te maken. Hierbij zijn Li2[Cd4(SPh)10] clusters toegepast als reactieve Cd bron, die gebruikt kan worden bij een relatief lage temperatuur. Deze eigenschap maakt Li2[Cd4(SPh)10] een veilige uitgangsstof die ook voor synthese op grote schaal gebruikt kan worden. De gevormde CdTe nanokristallen hebben een hoge kwantumopbrengst voor luminescentie (tot 37%) voor vertakte CdTe nanostructuren, die verhoogd kon worden tot 52% voor CdTe/CdSe kern-schil heterodeeltjes. De CdTe/CdSe nanokristallen zijn toegepast in een LED in combinatie met organische lagen voor elektronen- en gateninjectie. De diodes behaalden een efficiëntie van 0,35 cd/A. De eerste LEDs die lineair gepolariseerd licht uitzenden door gebruik te maken van uitgelijnde nanostaafjes worden beschreven in hoofdstuk 5. De lichtemitterende laag bestaat uit een dunne film van nanostaafjes met een aspectverhouding van 2,5, die uitgelijnd zijn door te wrijven met een fluwelen doek. In de diodes wordt de film met de uitgelijnde nanostaafjes geplaatst tussen twee organische lagen met elektronen- en gatengeleidende eigenschappen, om de injectie en efficiëntie te verbeteren. Op deze manier is een gepolariseerde LED verkregen met een emissie bij 620 nm, een luminescentie-efficiëntie van 0,65 cd/A en een externe kwantumopbrengst van 0,49%. De intensiteit van het gepolariseerde licht parallel aan de lengterichting van de staafjes is 1,5 keer zo hoog als in de loodrechte richting. De energieoverdracht in de aangeslagen toestand tussen bolvormige groenemitterende nanokristallen als donor en staafvormige roodemitterende nanokristallen als acceptor is beschreven in hoofdstuk 6. Voor deze toepassing zijn roodemitterende CdSe/CdS nanostaafjes met een hoge kwantumopbrengst voor luminescentie gemaakt en vervolgens gemengd met de groenemitterende bolvormige nanokristallen. In dit donor-acceptor systeem is de Förster-straal kleiner dan 6,6 nm. Deze afstand komt overeen met de som van de diameters van de bolvormige en cilindrische (staafvormige) nanokristallen. Hieruit volgt dat alleen de bolvormige nanokristallen die direct naast een nanostaafje liggen kunnen deelnemen in efficiënte energieoverdracht. Met behulp van een eenvoudige wrijftechniek is het mogelijk om de staafjes die gemengd zijn in een vaste film met

Samenvatting 119 bolvormige nanokristallen uit te lijnen. Deze uitgelijnde, gemengde films vertonen een gepolariseerde emissie, gecombineerd met een ongepolariseerde excitatie. De synthese van zeer dunne, gedoopte ZnSe:Mn nanodraden met een hoge luminescentie wordt beschreven in hoofdstuk 7. Deze synthese maakt gebruik van voorgevormde Li4[Zn10Se4(SPh)16] en Li2[Zn4(SPh)10] clusters samen met selenium en mangaanstearaat en vindt plaats bij milde temperatuur. De gevormde nanodraden zijn kristallijn en laten een heldere mangaan-foto-emissie zien. De diameter van de nanodraden kan gevarieerd worden tussen 1 en 3 nm, resulterend in een aspectverhouding van meer dan 80 voor nanodraden met een dikte van 2,5 nm. De emissie-eigenschappen zijn verder verbeterd door een CdSe schil te groeien op het ZnSe oppervlak. Dit leidt tot een kwantumopbrengst voor luminescentie tot 40%. De verhouding tussen bolvormige nanokristallen en nanodraden bij de reactie is te beïnvloeden door de seleniumconcentratie te variëren. Uitgelijnde nanodraden in een vloeistofstroom vertonen een zwakke gepolariseerde emissie, loodrecht op de lengterichting van de nanodraden. Samengevat, dit proefschrift beschrijft diverse syntheseroutes van halfgeleider nanokristallen die een hoge kwantumopbrengst voor luminescentie verbinden met grote controle over grootte, vorm en samenstelling van de deeltjes. De aantrekkelijke en karakteristieke optische eigenschappen van deze nanokristallen zijn bestudeerd, in combinatie met de energieoverdracht tussen de nanokristallen onderling en met geconjugeerde polymeren. Dit gaf de mogelijkheid om zowel nanofosforen als LEDs te maken, die dezelfde fascinerende optische eigenschappen bezitten als de nanokristallen, maar dan op een macroscopische schaal.

120 Samenvatting

121

List of Publications (2008) Highly luminescent ultra thin Mn doped ZnSe nanowires P. T. K. Chin, J. W. Stouwdam, R. A. J. Janssen in Preparation Synthesis and electroluminescence from branched CdTe nanocrystals P. T. K. Chin, J. W. Stouwdam, R. A. J. Janssen Nanotechnology 2008, 19, 205602 Energy transfer in hybrid quantum dot LEDs P. T. K. Chin, R. A. M. Hikmet, R. A. J. Janssen J. Appl. Phys. 2008, 104, 013108 (2007) Energy transfer and polarized emission in cadmium selenide nanocrystal solids with mixed dimensionality P. T. K. Chin, R. A. M. Hikmet, S. C. J. Meskers, R. A. J. Janssen Adv. Funct. Mater. 2007, 17, 3829 Highly luminescent CdTe/CdSe colloidal heteronanocrystals with temperature-dependent emission color P. T. K. Chin, C. de Mello Donegá, S. S. van Bavel, S. C. J. Meskers, N. A. J. M. Sommerdijk, R. A. J. Janssen J. Am. Chem. Soc. 2007, 129, 14880 Molecular imaging of macrophage activity in atherosclerotic plaques using bimodal PEG-micelles W. J. M. Mulder, G. J. Strijkers, K. C. Briley-Saboe, J. C. Frias, J. G. S. Aguinaldo, E. Vucic, V. Amirbekian, C. Tang, P. T. K. Chin, K. Nicolay, Z. A. Fayad, Mag. Res. Med. 2007, 58,1164 (2006) Quantum dots with a paramagnetic coating as a bimodal molecular imaging probe W. J. M. Mulder, R. Koole, R. J. Brandwijk, G. Storm, P. T. K. Chin, G. J. Strijkers, C. de Mello Donegá, K. Nicolay, A. Griffioen Nano Lett. 2006, 16, 1 Annexin A5 conjugated quantum dots with a paramagnetic lipidic coating for the mutimodal detection of apoptotic cells G. A. F. van Tilborg, W. J. M. Mulder, P. T. K. Chin, C. P. Reutelingsperger, K. Nicolay, G. J. Strijkers Bioconjug. Chem. 2006, 17, 865

122 (2005) Polarized light emitting quantum-rod diodes R. A. M. Hikmet, P. T. K. Chin, D. V. Talapin, H. Weller Adv. Mater. 2005, 17, 1436 (2004) The yeast phospholipid N-methyltransferases catalyzing the synthesis of phosphatidylcholine preferentially convert Di-C16:1 substrates both in vivo and in vitro. H. A. Boumann, P. T. K. Chin, A. J. R. Heck, B. de Kruijff, A. I. P. M. de Kroon J. Biol.Chem. 2004, 279, 40314

123

Curriculum vitae

Patrick Ted-Khong Chin werd geboren op 25 september 1976 te Voorschoten. Na het behalen van het mavodiploma aan het Aloysius College in Den Haag begon hij de opleiding voor chemisch laborant aan het MLO Haagland. Deze opleiding werd afgesloten met een stagejaar bij het Interfacultair Reactor Instituut in Delft, waar hij heeft gewerkt aan de accumulatie van technetium in spinazieplanten. Vervolgens is hij in september 1997 begonnen met de Hogere Laboratoriumopleiding Chemie aan de Hogeschool Rotterdam en Omstreken. Na het behalen van het propedeusediploma aan het HLO is hij in september 1998 begonnen met de studie Scheikunde aan de Universiteit Utrecht.

Deze studie werd in 2003 afgesloten met een afstudeeronderzoek uitgevoerd bij Philips Research in samenwerking met de vakgroep Condensed Matter and Interfaces van de Universiteit Utrecht. Dit afstudeeronderzoek was gericht op de toepassing van halfgeleider nanomaterialen in organische lichtgevende diodes en vormde de aanzet tot het latere promotieonderzoek. In september 2004 startte hij het promotieonderzoek dat beschreven staat in dit proefschrift bij de groep Moleculaire Materialen & Nanosystemen aan de Technische Universiteit Eindhoven onder de supervisie van prof.dr.ir. R. A. J. Janssen. Patrick Ted-Khong Chin was born on 25 September 1976. After obtaining the MAVO diploma at the Aloysius College in The Hague, he studied for chemical laboratory technician at the MLO Haagland. These studies were completed with a traineeship at the Interuniversity Reactor Institute in Delft, where he worked on the accumulation of technetium in spinach plants. Subsequently, he started in September 1997 studying Chemistry at the Hogeschool Rotterdam en Omstreken. After passing the first-year examination for this Higher Laboratory Education, he began in September 1998 studying Chemistry at the Utrecht University. The Master’s diploma was obtained in 2003 with a final research project at Philips Research in collaboration with the department Condensed Matter and Interfaces of Utrecht University. This research project focused on the application of semiconductor nanomaterials in organic light-emitting diodes and formed the basis for the subsequent PhD research project. In September 2004, he started with the PhD project described in this thesis at the Molecular Materials & Nanosystems group of the Eindhoven University of Technology under the supervision of prof.dr.ir. R. A. J. Janssen.

124

125

Dankwoord

Het is bijna zover, het einde van vier jaar onderzoek bij SMO/M2N. Deze tijd heeft mij een geweldige kans gegeven om met een ongekende hoeveelheid faciliteiten en vrijheid onderzoek te mogen doen in de halfgeleider nanomaterialen. Mede door de hulp en bijdragen van een groot aantal mensen van binnen de TU/e en daar buiten heeft dit tot dit proefschrift geleid.

Voor deze bijzondere tijd wil ik ten eerste mijn promotor René Janssen bedanken. De grote vrijheid, begeleiding en toch kritische kijk op de zaken waren zeer waardevol voor mijn onderzoek.

Ook wil ik mijn begeleider van mijn afstudeerproject, Rifat Hikmet, bedanken. Zonder jou was ik waarschijnlijk niet in dit onderzoek terecht gekomen. Jouw onderzoek naar QD-LEDs was voor mij het begin van deze fascinerende tijd in het QD onderzoek. De vele hulp en discussies voor en tijdens mijn promotietijd zijn van zeer grote waarde geweest.

Als de optische spectra weer eens onbegrijpelijke pieken en dalen vertoonden had jij er bijna altijd wel een heldere verklaring voor. Stefan, jouw spectroscopische kennis en creativiteit maakten de meest onmogelijke ideeën mogelijk. Celso jouw hulp en discussies hebben heel veel bijgedragen aan mijn onderzoek. Voor dat jij me liet zien hoe je CdSe QDs moest bakken had ik nauwelijks een idee hoe QDs gemaakt werden. Ik ben je zeer dankbaar voor onze succesvolle samenwerking die we bij mijn volgende baan hopelijk een vervolg kunnen geven.

Wiljan jij bent een top bollenbakker, jouw vernieuwende ideeën gecombineerd met mijn wilde plannen hebben uiteindelijk tot hele leuke resultaten geleid.

Elektronenmicroscopie is een essentieel onderdeel in dit proefschrift maar ook een studie apart. Svetlana de vele tijd die het jou kostte om mij te leren hoe ik de microscoop moest bedienen, is voor mij van onschatbare waarde geweest. Nico bedankt voor de hulp met de Titan. Dit heeft echte supermooie foto’s opgeleverd. Joachim, Paul en Niek bedankt voor alle hulp en adviezen.

126

Ton van den Biggelaar, ondanks jouw drukke werkschema had je altijd tijd me te helpen met het maken van devices en om de opdamper te redden als er weer eens iets fout ging. Heel erg bedankt voor deze prettige samenwerking.

Willem Mulder, jij bent echt een MRI-contrast-topper, jouw lipide-gecoate QDs hebben mij zo enthousiast gemaakt voor de nano/bio combi dat ik er nu zelf ook in ga werken. Ik hoop dat we in de toekomst weer een leuke samenwerking kunnen starten.

Harm, jouw wil ik bedanken voor je inzet en interesse, je hebt in een zeer korte tijd

veel mooie resultaten laten zien.

Erik en Lars, jullie wil ik bedanken voor onze samenwerking. Hopelijk kunnen we nog wat mooie resultaten scoren.

Mevr. Elemans-Mehring uw ICP element analyses waren zeer verhelderend en

belangrijk om de syntheses te begrijpen, bedankt.

Frank bedankt voor de vele hulp met de glovebox, ESR metingen, onze trip naar de MRS samen met Arjan en Simon. Arjan .. en Simon, bedankt voor deze US and A ervaring.

Verder wil ik alle mensen van M2N en Lab 3 bedanken voor alle hulp, leuke conferenties en werkomgeving en natuurlijk speciaal voor de mooie Lab3 muziek Jeroen, Maarten P. en Bas.

Natuurlijk wil ik ook mijn kamergenotengenoten Theresa, René, Ingrid, Martijn, Jan en Thomas bedanken voor de plezierige tijd en voor het feit dat ik jullie 4 jaar tegen een enorme berg troep op mijn bureau heb laten kijken. Verder wil ik ook mijn broers Roy en Ron bedanken. Onze gezamenlijke interesse in de techniek zijn altijd een belangrijke stimulans geweest. Roy, jouw wiskundige en natuurkundige hulp waren de laatste zetjes naar een succesvolle universitaire studie. Tot slot wil mijn ouders bedanken voor al hun hulp en steun. Zonder jullie hulp was ik nooit zo ver gekomen. Martine, het laatste plekje in dit proefschrift is voor jou. Ik vind het echt super dat je bijna al mijn teksten en publicaties corrigeerde. Zonder jou was dit nooit zo’n mooi boekje geworden. Martine, ik hou van jou ……….