Lead Chalcogenide Quantum Dots and Quantum Dot ... - OPUS 4 [PDF]

Photodetectors fabricated with PbSe-MoS2 hybrids on rigid and flexible substrates show a long-time air-stable. NIR photo

12 downloads 40 Views

Recommend Stories


Quantum Dots
Forget safety. Live where you fear to live. Destroy your reputation. Be notorious. Rumi

Quantum Dots
You often feel tired, not because you've done too much, but because you've done too little of what sparks

quantum dot
If you feel beautiful, then you are. Even if you don't, you still are. Terri Guillemets

InP quantum dots and quantum dashes
Everything in the universe is within you. Ask all from yourself. Rumi

ZnS core-shell quantum dots and PbS quantum dots into
The wound is the place where the Light enters you. Rumi

GaAs quantum dots
We must be willing to let go of the life we have planned, so as to have the life that is waiting for

SiGe quantum dots
Your task is not to seek for love, but merely to seek and find all the barriers within yourself that

GaAs quantum dots
When you do things from your soul, you feel a river moving in you, a joy. Rumi

Shell Tunable Quantum Dots
The only limits you see are the ones you impose on yourself. Dr. Wayne Dyer

Si quantum dots
Life is not meant to be easy, my child; but take courage: it can be delightful. George Bernard Shaw

Idea Transcript


.

Lead Chalcogenide Quantum Dots and Quantum Dot Hybrids for Optoelectronic Devices

Bleichalkogenidquantenpunkte und Quantenpunkthybridverbindungen f¨ ur optoelektronische Bauteile

Der Technischen Fakult¨ at der Friedrich-Alexander-Universit¨ at Erlangen-N¨ urnberg zur Erlangung des Doktorgrades Dr.-Ing.

vorgelegt von

Julia Schornbaum aus Schwabach

.

Als Dissertation genehmigt von der Technischen Fakult¨ at der Friedrich-Alexander-Universit¨ at Erlangen-N¨ urnberg Tag der m¨ undlichen Pr¨ ufung: 18.12.2015 Vorsitzender des Promotionsorgans: Prof. Dr. Peter Greil 1. Gutachterin: Prof. Dr. Jana Zaumseil 2. Gutachter: Prof. Dr. rer. nat. Alexander Eychm¨ uller 3. Gutachter: Prof. Dr. rer. nat. Erdmann Spiecker

Abstract Semiconductor quantum dots (QDs) exhibit remarkable properties, which include a size-tunable band gap and narrow emission bands. They are also suitable for largearea and low-cost fabrication, due to their solution-processability. Consequently, QDs are very promising for future applications in printable optoelectronic devices. Near-infrared (NIR) active lead chalcogenide QDs hold an enormous potential, as they exhibit optical properties in a wavelength regime, where efficient photoactive materials are rare. In order for QDs to become commercially viable in future optoelectronics, the current performance of QD devices still requires improvement. Therefore, it is essential to optimize and to understand charge carrier transport in QD thin-films. This thesis introduces new strategies to synthesize lead chalcogenide QD hybrids in order to improve charge transport. Moreover, it describes the fabrication and characterization of PbS QD light-emitting field-effect transistors (LEFETs), which enable the investigation of charge carrier transport and recombination dynamics. Hybrids were synthesized with a one-pot hot-injection synthesis, where PbSe QDs were directly and without any linker molecules grown on the supporting materials, such as single-walled carbon nanotube (SWNT) bundles, few-layered graphene (FLG) flakes, or transition metal dichalcogenide (TMD) flakes (i.e., MoS2 and WS2 ). These hybrid materials combine efficient NIR light absorption (realized by the QDs) and high charge carrier mobility (realized by the respective supporting material). PbSe QDs grown in the presence of SWNTs form half-ring shaped dots around the nanotube bundles and exhibit a preferred orientation of their {002} lattice planes perpendicular to the nanotube bundles. PbSe-MoS2 hybrids reveal a preferred orientation of the QDs on the MoS2 nanoflakes, inferring an epitaxial growth of the QDs on the MoS2 nanoflakes. Photodetectors fabricated with PbSe-MoS2 hybrids on rigid and flexible substrates show a long-time air-stable NIR photoresponse. PbS QD LEFETs were realized with electrolyte-gated, ligand-exchanged QD thin-films. These LEFETs are the first LEFETs that have ever been demonstrated with a zero-dimensional material. Electrolyte-gating enabled high charge carrier densities and, thus, ambipolar transistor characteristics were measured and NIR light emission was detected. The LEFETs show a significant enhancement of external quantum efficiency, photoluminescence intensity, and photoluminescence average lifetime with charge carrier density. The increased emission efficiency at high charge carrier density is a result of an effective deactivation of non-radiative decay channels and a subsequent dominant trion emission.

Zusammenfassung Halbleiterquantenpunkte (QDs) besitzen außergew¨ohnliche Eigenschaften, wie eine gr¨oßenabh¨angige Bandl¨ ucke und schmale Emissionsbanden. Außerdem sind QDs f¨ ur eine großfl¨achige und kosteng¨ unstige Fertigung geeignet, da sie in L¨osung prozessiert werden k¨onnen. Durch diese Eigenschaften ist der zuk¨ unftige Einsatz von QDs in druckbaren optoelektronischen Bauteilen sehr vielversprechend. Besonders Bleichalkogenid-QDs besitzen ein enormes Potential f¨ ur diese Anwendungen, da ihre optischen Eigenschaften im nahinfraroten Spektralbereich liegen, in dem es sonst nur wenige effiziente photoaktive Materialien gibt. Damit QDs in Zukunft in optoelektronischen Bauteilen kommerziell genutzt werden k¨onnen, muss die derzeitige Effizienz der QD-basierten Bauteile noch verbessert werden. F¨ ur eine solche Verbesserung ist es essentiell den Ladungstr¨agertransport in d¨ unnen Filmen aus QDs zu verstehen und zu optimieren. Diese Arbeit besch¨aftigt sich mit neuen Strategien zur Herstellung von PbSe QD-Hybridverbindungen, um Verbesserungen im Ladungstr¨agertransport zu erreichen. Des Weiteren beschreibt sie die Herstellung und Charakterisierung von lichtemittierenden Feldeffekttransistoren (LEFETs) mit PbS QDs als Halbleitermaterial. Diese Transistoren erm¨oglichen die Untersuchung des Ladungstr¨agertransports und der Rekombinationsdynamiken in PbS QDs. Die PbSe QD-Hybridverbindungen wurden nasschemisch mit der sogenannten hot-injection Technik synthetisiert. PbSe QDs wurden direkt und nicht-kovalent auf dem jeweiligen Tr¨agermaterial, n¨amlich auf einwandigen Kohlenstoffnanor¨ohrchen ¨ (SWNTs), mehrlagigen Graphenfl¨ockchen oder Fl¨ockchen der Ubergangsmetalldichalkogenide MoS2 und WS2 gewachsen. Diese Hybridmaterialien kombinieren eine effiziente Absorption von nahinfrarotem Licht (realisiert durch die Nanopartikel) mit hohen Ladungstr¨agermobilit¨aten (realisiert durch das jeweilige Tr¨agermaterial). Die PbSe QDs, die in Gegenwart von SWNTs entstanden sind, besitzen die Form von Halbringen, da sie um die SWNT B¨ undel herum gewachsen sind. Außerdem ist eine Mehrzahl der QDs mit ihren {002} Netzebenen senkrecht zu den SWNT B¨ undeln orientiert. Bez¨ uglich der PbSe-MoS2 Hybridverbindungen konnte festgestellt werden, dass die QDs eine bevorzugte Orientierung auf den MoS2 Nanofl¨ockchen aufweisen. Dies deutet auf ein epitaktisches Wachstum der QDs auf den MoS2 Nanofl¨ockchen hin. Anschließend wurden PbSe-MoS2 Photodetektoren auf ihre Lichtempfindlichkeit und Stabilit¨at hin untersucht. Es konnte gezeigt werden, dass diese Photodetektoren u ¨ ber einen langen Zeitraum und u ¨ ber viele Schaltzyklen an Luft stabil sind. Außerdem ergaben die Experimente, dass flexible PbSe-MoS2 Photodetektoren eine hohe Stabilit¨at gegen¨ uber extremen Biegeradien besitzen. LEFETs wurden mit d¨ unnen Filmen aus PbS QDs hergestellt und u ¨ ber ein elektrolytisches Dielektrikum gesteuert. Um die d¨ unnen PbS QD-Filme elektrisch leitf¨ahig zu machen, wurde ein Ligandenaustausch durchgef¨ uhrt. Die

hergestellten Transistoren sind die ersten LEFETs die jemals mit einem nulldimensionalen Material realisiert werden konnten. Die Verwendung eines Elektrolyts anstelle eines konventionellen Dielektrikums erlaubt die Akkumulation von hohen Ladungstr¨agerdichten in den PbS QD-Filmen. Infolgedessen weisen die Transistoren ambipolares Verhalten auf und nahinfrarote Emission konnte detektiert werden. Die LEFETs zeigen eine deutliche Erh¨ohung der externen Quanteneffizienz, der Photolumineszenz-Intensit¨at und der durchschnittlichen PhotolumineszenzLebensdauer mit steigender Ladungstr¨agerdichte. Diese erh¨ohte Emissionseffizienz bei hohen Ladungstr¨agerdichten resultiert aus einer effektiven Deaktivierung von nichtstrahlenden Zerfallskan¨alen und einer daraus folgenden dominanten Trionemission.

Contents 1 Introduction 2 Background 2.1 Colloidal Semiconductor Quantum Dots . . . . . . . . . 2.1.1 Hot-injection Synthesis in an Organic Medium . . 2.1.2 Optical and Electrical Properties . . . . . . . . . 2.1.3 Quantum Dot Solids . . . . . . . . . . . . . . . . 2.1.4 Lead Chalcogenide Quantum Dots . . . . . . . . 2.2 Carbon Nanotubes . . . . . . . . . . . . . . . . . . . . . 2.2.1 Fabrication and Processing . . . . . . . . . . . . . 2.2.2 Optical and Electrical Properties . . . . . . . . . 2.3 Layered Materials . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Graphene . . . . . . . . . . . . . . . . . . . . . . 2.3.2 Transition Metal Dichalcogenides . . . . . . . . . 2.4 Semiconductor Quantum Dot Hybrids . . . . . . . . . . . 2.4.1 Quantum Dots-Carbon Nanotubes . . . . . . . . 2.4.2 Quantum Dots-Graphene . . . . . . . . . . . . . . 2.4.3 Quantum Dots-Transition Metal Dichalcogenides 2.5 Solution-Processed Quantum Dot Optoelectronics . . . . 2.5.1 Photodetectors . . . . . . . . . . . . . . . . . . . 2.5.2 Field-effect Transistors . . . . . . . . . . . . . . .

1

. . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

5 6 8 12 15 20 22 23 25 27 27 30 32 34 35 37 38 39 45

3 Experimental Part 3.1 Chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Syntheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 PbSe Quantum Dots . . . . . . . . . . . . . . . . . . . . 3.2.2 PbS Quantum Dots . . . . . . . . . . . . . . . . . . . . . 3.2.3 PbSe-Single-walled Carbon Nanotube Hybrids . . . . . . 3.2.4 PbSe-Layered Material Hybrids . . . . . . . . . . . . . . 3.3 Device Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 PbSe Quantum Dot Hybrid Photodetectors . . . . . . . 3.3.2 PbS Quantum Dot Light-emitting Field-effect Transistors 3.4 Material Characterization . . . . . . . . . . . . . . . . . . . . . 3.4.1 Absorption Spectroscopy . . . . . . . . . . . . . . . . . . 3.4.2 Photoluminescence Spectroscopy . . . . . . . . . . . . . 3.4.3 Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . 3.4.4 Fourier-transform Infrared Spectroscopy . . . . . . . . . 3.4.5 X-ray Photoelectron Spectroscopy . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

57 58 58 58 60 61 62 62 63 63 66 66 67 72 73 74

. . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

VII

3.5

3.4.6 Scanning Electron Microscopy . . . . . . . 3.4.7 Transmission Electron Microscopy . . . . . Device Characterization . . . . . . . . . . . . . . 3.5.1 Photoresponse Measurements . . . . . . . 3.5.2 Light-emitting Transistor Characterization

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

4 Colloidal PbSe Quantum Dot Hybrid Materials 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids . . . . . . . . 4.2.1 Fabrication of PbSe-SWNT Hybrids . . . . . . . . . . . 4.2.2 Characterization of PbSe-SWNT Hybrids . . . . . . . . 4.2.3 PbSe QD Size and Shape Optimization . . . . . . . . . 4.2.4 Three-dimensional Morphology . . . . . . . . . . . . . 4.2.5 Growth Mechanism . . . . . . . . . . . . . . . . . . . . 4.2.6 PbSe-SWNT Photodetectors . . . . . . . . . . . . . . . 4.2.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 PbSe Quantum Dot-Layered Material Hybrids . . . . . . . . . 4.3.1 Fabrication of PbSe-Layered Material Hybrids . . . . . 4.3.2 Nanoflake Thicknesses of Layered Materials . . . . . . 4.3.3 PbSe-MoS2 Interface Characterization . . . . . . . . . 4.3.4 Control Experiments with Ex Situ Fabricated Hybrids 4.3.5 PbSe-MoS2 Interface Visualization . . . . . . . . . . . 4.3.6 Ligand Arrangement . . . . . . . . . . . . . . . . . . . 4.3.7 Growth Mechanism . . . . . . . . . . . . . . . . . . . . 4.3.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 PbSe-MoS2 Photodetectors . . . . . . . . . . . . . . . . . . . . 4.4.1 Basic Photodetector Characterization . . . . . . . . . . 4.4.2 Repeated Switching Stability . . . . . . . . . . . . . . 4.4.3 Dark Current . . . . . . . . . . . . . . . . . . . . . . . 4.4.4 Photoresponse Time . . . . . . . . . . . . . . . . . . . 4.4.5 Applied Bias Dependence . . . . . . . . . . . . . . . . 4.4.6 Photoresponse Air-stability . . . . . . . . . . . . . . . 4.4.7 Control Devices with Ex Situ Fabricated Hybrids . . . 4.4.8 Flexible Photodetectors . . . . . . . . . . . . . . . . . 4.4.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Conclusion of Chapter 4 . . . . . . . . . . . . . . . . . . . . . 5 Quantum Dot Light-emitting Field-effect Transistors 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 PbS Quantum Dot Light-emitting Field-effect Transistors 5.2.1 PbS Quantum Dots . . . . . . . . . . . . . . . . . 5.2.2 Quantum Dot Thin-films . . . . . . . . . . . . . . 5.2.3 Electrolyte-gated Thin-film Transistors . . . . . . 5.2.4 Electroluminescence . . . . . . . . . . . . . . . . 5.2.5 External Quantum Efficiency . . . . . . . . . . . 5.2.6 Top-gated Thin-film Transistors . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . .

. . . . .

74 75 79 79 79

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

83 84 85 85 86 90 92 93 97 98 99 99 100 103 107 108 109 110 112 113 113 115 115 116 117 118 119 119 121 121

. . . . . . . .

123 124 125 125 126 130 133 137 139

5.3

5.4

5.2.7 Summary . . . . . . . . . . . . . . . . . . . . . . Charge Carrier Dynamics in PbS Quantum Dot LEFETs 5.3.1 Gate Voltage Dependence of Photoluminescence . 5.3.2 Transient Photoluminescence Analysis . . . . . . 5.3.3 Proposed Model of Recombination Channels . . . 5.3.4 Summary . . . . . . . . . . . . . . . . . . . . . . Conclusion of Chapter 5 . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

141 142 142 142 148 151 152

6 Conclusion and Outlook

153

Bibliography

157

Appendix

203

A List of Abbreviations

205

B List of Symbols

209

C Publications

213

D Acknowledgments

215

1 Introduction Our daily life cannot be imagined without optoelectronic devices, used for fast data communication, energy generation, brilliant displays, and home-illumination concepts. Optoelectronic devices either convert an optical signal into an electrical signal or vice versa. Therefore it is very important to develop both, new light-emitting solutions and innovative methods to detect light. To ensure well performing optoelectronic devices, an efficient integration of the optical active material into the device is fundamental. The integration of conventional bulk materials often needs expensive processing and packaging techniques and is not compatible with flexible substrates. For this reason, trends are going towards solution-processable materials. Their simple processing techniques, which include printing, spraying, coating, or casting, are cheap and easy to scale-up.[1–10] Moreover, solution-processed materials can be used to fabricate electronic devices on flexible substrates. However, the integration of solution-processable materials in optoelectronic applications requires innovative solutions both for materials and technologies and, thus, current research is dealing with a broad range of new concepts. Recently, inorganic quantum dots (QDs) were used to improve the color performance of liquid crystal display televisions, which depicts the commercial breakthrough for QDs in optoelectronics.[11] The size of QDs is smaller than the exciton Bohr radius, resulting in quantum confinement in three-dimensions.[12–14] This confinement leads to a size dependent band gap and enables a precise and systematic wavelength tunability.[14] QDs exhibit a spectrally narrow emission, creating brilliant colors and very high detection sensitivities, while they additionally possess the advantage of solution-processability. Therefore, QDs can be integrated in almost every device, guaranteeing a reduction in manufacturing costs and enabling large-area processing.[15,16] Depending on the material and the size, QDs not only reveal their outstanding properties in the visible, they also show excellent performance in the near-infrared (NIR) wavelength regime, from 750 to 2400 nm.[17] Infrared light-emitting sources and infrared detectors are important for quality control in pharmaceutical and food industry,[18–20] tumor detection in biological imaging,[21–23] night vision, and

1

1 Introduction optical data communication by means of optical fibers.[21] The standard data communication wavelength ranges from around 1300 to 1700 nm, with the most important band at 1550 nm,[24] while some living tissue transparent windows, which are essential for deep tissue imaging, range between 840 to 1680 nm.[21] Only a few polymers and organic dyes can cover these wavelength ranges, but their efficiencies are low. QDs exhibit very high emission and detection efficiencies in the NIR and their size can be tuned to the desired wavelength. These properties render QDs superior to other materials for NIR applications.[25,26] Narrow band gap lead chalcogenide QDs are NIR active with very high photoluminescence (PL) quantum yields (QYs).[27] PbX (X = S, Se, Te) QDs exhibit nearly identical and large Bohr radii for electrons and holes, therefore they are able to simultaneously transport electrons and holes (ambipolar).[14,28] However, this transport of charges is still very low in PbX QD thin-films, limiting their use in commercial applications. In order to make QD materials attractive to industry, it is important to find new solutions to improve the charge carrier transport and to understand charge carrier dynamics. The following thesis addresses both of these topics, discussing lead chalcogenide QDs and QD hybrids for solution-processable optoelectronic devices. Chapter 2 gives background information, which is necessary to better understand the experimental work. First, QD synthesis, properties, and applications are discussed, followed by an introduction of carbon nanotubes (CNTs) and layered materials, namely graphite/graphene and transition metal dichalcogenides (TMDs). Then, an overview about semiconductor QD hybrids is given, introducing different coupling strategies and hybrid properties. Chapter 2 concludes with a discussion about semiconductor QD optoelectronics, focusing on photodetectors and (lightemitting) field-effect transistors ((LE)FETs). The third chapter introduces all the experimental procedures that were used for material preparation, device fabrication, and material and device characterization. It first describes syntheses parameters and subsequently discusses photodetector and transistor fabrication details. Then, the theoretical background of all characterization methods is shortly explained, and concrete characterization parameters, equipment, and calculations are named. Finally, Chapter 3 comments on parameters, equipment, and calculations that were applied to extract figures of merit from photodetector and LEFET measurements. The first results section of this thesis starts with Chapter 4. It covers new syntheses strategies to produce lead selenide QD hybrids as well as a discussion on their application in NIR sensitive photodetectors. The innovative hybrid ma-

2

terials combine efficient NIR light absorption (realized by the QDs) and a high charge carrier mobility (realized by the supporting material), thus improving the photodetector performance compared to detectors based on QDs only. In an in situ hot-injection syntheses, PbSe QDs are directly and non-covalently attached to the supporting materials, such as untreated single-walled carbon nanotubes, few-layer graphene sheets, or transition metal dichalcogenide layers (MoS2 or WS2 ). Subsequently, the growth mechanism was studied through a systematic change of synthesis parameters and an extended analysis with transmission electron microscopy techniques, including electron tomography, selected-area electron diffraction, and high-resolution transmission electron microscopy. Finally, PbSeMoS2 hybrids were used to fabricate air-stable NIR sensitive photodetectors on rigid glass and flexible polymer substrates. In order to improve efficiencies of QD-based optoelectronic devices, it is very important to understand and control charge carrier dynamics at high charge carrier densities. Ambipolar LEFETs provide an ideal system to study these processes, however until now LEFETs were only achieved with one-dimensional (e.g., CNTs), two-dimensional (e.g., MoS2 ), and bulk (e.g., polymers) materials, but no LEFETs with zero-dimensional materials could be realized. Chapter 5, featuring the second part of the result section, demonstrates the fabrication and characterization of the first LEFET based on a zero-dimensional material. These LEFETs were based on electrolyte-gated ligand-exchanged PbS QD thin-films. Electrolyte-gating allows to use the transistor in the ambipolar regime, thus achieving NIR light emission from a confined region within the channel. The tunability of the spectral position of the electroluminescence was investigated by using different sized PbS QDs for thin-film preparation. A significant enhancement of external quantum efficiencies was observed and further characterized using time-resolved PL experiments. At the end, the Conclusion summarizes the results, talks about their scientific impact, and makes suggestions for further experiments.

3

2 Background

This chapter provides all the information, which is necessary to understand the experiments and results of this thesis. It starts with an introduction about quantum dot synthesis, properties, and application. This is followed by a discussion about carbon nanotubes and layered materials, namely transition metal dichalcogenides and graphite/graphene. The next part deals with coupling strategies of quantum dots and either one-dimensional carbon nanotubes or two-dimensional layered materials and discusses the properties of these hybrid structures. The last section is about quantum dot-based photodetectors and (light-emitting) field-effect transistors.

2 Background

2.1 Colloidal Semiconductor Quantum Dots In the early 1980s three independent research groups around the world, Efros et al.[29–31] in the Soviet Union, Brus et al.[32–34] in the United States, and Henglein et al.[35] in Germany discovered color changes in the absorption spectra of semiconductor nanoparticles depending on their size. This phenomenon started an enormous interest in the research field of semiconductor nanoparticles. Unlike bulk materials, where the composition of the material defines its optoelectronic properties, the optical and electrical properties of a nanomaterial can be tailored by its shape and its physical dimensions. The most important characteristic of a semiconductor is its band gap Eg . A conventional semiconductor exhibits a distinct and material-dependent band gap. If light with a photon energy larger than the band gap hits a semiconductor with a direct band gap, electrons will be excited from the valence band (VB) to the conduction band (CB). If the electron and the hole are bound by Coulombic interactions, they form an electron-hole pair, which is named exciton. Semiconductor particles are called quantum dots (QDs), when the size of the particle is smaller than the exciton Bohr radius αB , defined as αB =

ε0 εh2 , πµrm e2

(2.1)

where ε0 and ε are the permittivity of the free vacuum and the relative permittivity mh of the semiconductor, µrm is the reduced mass of the electron and hole ( (mmee+m ), h) [12] and e is the electron charge. Such a small particle size leads to a quantum confinement of electrons and holes in all three dimensions.[13,14] Due to the reduced dimensions, QDs are considered as zero-dimensional materials. Figure 2.1 shows that the band diagram changes progressively from bulk continuous energy bands into molecule like discrete energy levels, with QDs somewhere in an intermediate state. Therefore, the band gap of QDs increases with decreasing QD size.[14,36,37] This effect is explained by the particle-in-a-sphere model.[30,38] The semiconducting QD is considered as a spherical mass, which is surrounded by an insulating barrier. The three-dimensional potential wall confines all photoexcited electrons and holes inside the spherical QD of constant potential. This scenario of a particle of mass m confined in a spherical symmetric potential wall V , can be written as:

V (r) =

 0,

if r < a

∞, if r ≥ a

6

,

(2.2)

2.1 Colloidal Semiconductor Quantum Dots

Figure 2.1: Energy band diagram for atoms, clusters, particles, and bulk materials. With an increasing number of atoms, the spacing between the discrete energy states decreases approaching a continuous band diagram for bulk materials. The energy band diagram of QDs can be found somewhere in an intermediate state. Figure adapted from Wheeler et al.[12]

where a defines the radius of the sphere. Semiconducting QDs exhibit a crystal lattice, which describes a periodic potential, therefore, the wavefunctions of electrons and holes can be expressed as Bloch functions. It is possible to perform an effective mass approximation for the VB and the CB. Subsequently, solving the Schr¨odinger equation and taking into account that the wavefunction has to drop to zero at the QD edge (considered by β in equation 2.3), leads to discrete allowed energy levels for the particle. This is why the absorption and emission bands in QDs are discrete and not continuous as in bulk semiconductors. Figure 2.2 displays the density of states (DOS) for bulk, two-dimensional, one-dimensional, and zero-dimensional semiconductors. It can be seen that the number of states in a given energy interval depend on the dimensionality of the semiconductor, leading to a set of discrete delta functions for zero-dimensional QDs.[38]

0D v

v v v v

v

DOS(E)

1D DOS(E)

2D DOS(E)

DOS(E)

3D

v

v

Energy

Energy

Energy

Energy

Figure 2.2: DOS for different dimensional semiconductors. In bulk materials (3D) DOS are proportional to the square root of the energy, while in two-dimensional materials DOS are independent of the energy. In one-dimensional semiconductors DOS are proportional to one divided by the square root of the energy. In zero-dimensional structures, like QDs, DOS are discrete, described by a set of delta functions. Figure adapted from reference.[38]

7

2 Background Moreover equation 2.3 shows that the QD band gap exhibits an inverse squared dependence on the radius rQD , explaining the blueshift of the QD absorption with decreasing QD size:[30,38] h2 k 2 }2 k 2 }2 β 2 , Eg = 2 ∗ = = 2 8π m 2m∗ 2m∗ rQD

(2.3)

where Eg is the energy of the band gap, h is the Planck’s constant, } is the reduced Planck’s constant, and k the wave vector defined as β/rQD . The term colloidal quantum dots (CQDs) is commonly used for QDs synthesized in solution in order to differentiate them from QDs prepared by deposition techniques. However, within the scope of this thesis, QDs were always synthesized in solution, therefore the term QDs always stands for CQDs. Solution synthesized QDs exhibit an inorganic crystalline core stabilized by organic or inorganic surfactants (Figure 2.3(a)). QDs are usually between 2 and 50 nm in size, where the smallest QDs exhibit less than 100 atoms and the largest QDs consist of more than 100.000 atoms. QDs have a very large surface-to-volume ratio, with a ranging percentage of surface atoms between > 75 % and < 0.5 %,[39] this is why the surface has an enormous influence on QD properties, including optical performance and charge transport.[40]

Figure 2.3: (a) Schematic illustration of a PbSe QD. The outer shell of the QD is made of lead atoms protected by surfactants. (b) TEM image of a PbSe QD with clearly resolved lattice planes.

2.1.1 Hot-injection Synthesis in an Organic Medium In the beginning of QD research it was extremely difficult to synthesize monodisperse QDs in a controlled and reproducible way, with clearly assessable physical and chemical properties. The pioneering work by Brus and Henglein described fundamental synthesis routes and a principle understanding of the electrical prop-

8

2.1 Colloidal Semiconductor Quantum Dots erties. However, the first syntheses lead to polydisperse samples with highly defect rich QDs.[33,34,41–44] Therefore, an optimized synthesis was necessary to produce high quality QDs, with a narrow size distribution and a good crystallinity. A tight control on QD size, shape, and surface chemistry assures a controlled tunability of the physical and chemical properties of the QDs.[45] In 1993 Murray, Norris, and Bawendi introduced the hot-injection synthesis, a milestone in QD research, leading to an immense increase in QD interest.[46] The hot-injection process is a very versatile synthesis forming monodisperse, highly luminescent, well-passivated and almost defect free QDs by a temporally discrete nucleation. Murray et al. synthesized cadmium chalcogenide QDs in the size range of 1.2-11.5 nm by using organometallic Cd and chalcogenide (Se, S, or Te) precursors and a high-boiling point, nonpolar coordinating solvent.[39,46] In principal, the typical hot-injection synthesis system for QDs consists of three components: the precursors, which determine the components of the inorganic core, the surfactant, which ensures the colloidal stability, prevents aggregation, passivates unsaturated surface dangling bonds, and controls nucleation and growth and the solvent, which provides the synthesis medium. In the early work, the coordinating solvent was also used as surfactant.[39,46] Later on, Yu et al. established a hotinjection synthesis with a non-coordinating solvent, i.e., octadecene, together with coordinating ligands. Compared to syntheses with a coordinating solvent, syntheses with non-coordinating solvents and separately coordinating surfactants improve the flexibility and tunability of the syntheses, because the size and size distribution of the QDs can be freely controlled by the surfactant concentration.[47] Nowadays, size, size dispersion, morphology, and composition of QDs prepared by a hot-injection synthesis are controllable by manipulating many synthetic parameters, like surfactants and precursors (materials as well as concentration), injection temperature, growth temperature, and growth time.[48,49] Nucleation and growth in solution phase. The nucleation and early growth of QDs follows the basic concepts of the LaMer theory, which was established in 1950.[50] This theory claims that it is necessary to temporally separate nucleation and growth in order to form monodisperse particles. After the rapid injection of a precursor at room temperature into a hot liquid, a burst of nucleation generates many nuclei, followed by the growth of the existing nuclei.[39] However, the LaMer nucleation theory cannot explain all phenomena observed during QD growth. Therefore Rempel et al.[51,52] developed a 5 step process. Monomers are generated from the organometallic precursors leading to an increase of particles (first step)

9

2 Background and a subsequent formation of small clusters (second step). The formation of these clusters is followed by a size-focusing process (third step), a pseudo steady state with an invariant mean size and size distribution (fourth step), and finally, Ostwald ripening, leading to defocusing (fifth step). The defocusing during the Ostwald ripening process comes from a dissolution of small crystals with high surface energy at the expense of a further growth of large crystals with low surface energy (Figure 2.4).[53] Combining the existing theories, QD growth can be described in the following manner: During nucleation the precursor decomposes at relatively high temperatures creating a supersaturation of monomers in solution. Then, nucleation and growth are temporally separated by a rapid injection of the second precursor. The consequent burst of nucleation lowers the monomer concentration below the nucleation threshold and produces many QD nuclei. At this point, monomers are still in solution, but they will only add to already existing nuclei and not form new ones. The rapid injection of the second precursor, which is at room temperature, into the hot solution with the first precursor induces a drastic temperature drop. This temperature drop ensures a lower growth temperature than injection temperature and therefore prevents a further nucleation during the growth stage. The growth of the existing nuclei happens by an incorporation of monomers.[49] As long as the monomer concentration in solution is high, the nuclei growth is diffusion-limited, which results in a size-focusing. Due to the larger surface to volume ratio of small nuclei compared to large ones, small nuclei will grow faster than large nuclei, leading to an adaption of all QD sizes.[48,54] If one or more precursors are fully consumed, slow Ostwald ripening occurs, leading to defocusing.[53] This reversed size-focusing leads to a loss of monodispersity and a consequently larger size distribution. During focusing the number of growing QDs keeps constant while it decreases during defocusing, whereas the number of monomers drastically decreases during focusing and remains unchanged during defocusing.[39,55]

Figure 2.4: Schematic illustration of the Ostwald ripening process. Small crystals dissolve due to their high surface energies and are subsequently integrated into large crystals.

10

2.1 Colloidal Semiconductor Quantum Dots Function and influence of surfactants. Usually, long-chain carboxylic acids, phosphonic acids, or amines are used as surfactants during QD synthesis. These surfactants are made of an anchoring group (e.g., amino, carboxylic, phosphonato etc.), which has a high affinity to the QD surface and an alkyl chain, typically with twelve to eighteen carbon atoms.[15,56] The surfactants coordinate dynamically to the surface of the growing QDs, thus forming a steric barrier for reactants. Consequently, the QD growth is relatively slow at high temperatures, which results in a self-annealing of the QDs. During the self-annealing process internal grain boundaries and defects are eliminated, leading to highly crystalline and defect-free QDs.[49] Purification process. After the synthesis, the QDs have to be separated from the growth solution in order to get rid of unreacted precursors, side-products, and excess surfactants. This post-synthesis purification process is done by adding polar non-solvents to the growth solution and thus, destabilizing the colloidal suspension, leading to QD aggregation and precipitation. Subsequently, to form a stable and purified colloidal suspension, the QDs are re-dispersed in a nonpolar solvent.[39,49] Through the stepwise addition of a polar solvent the colloidal stability of the QDs is gradually reduced. Large QDs tend to aggregate and precipitate before the small ones, due to stronger attractive forces between large QDs than between small ones. This size-selective precipitation enables to separate QD fractions with very narrow size distributions.[39,46,48] Although the purification process is absolutely necessary, the effects of this post-synthesis step on the final QD properties (e.g., emission efficiencies) should not be neglected.[57,58] Controlled growth of anisotropic nanocrystals. The physical and chemical properties of QDs not only depend on their size, but also on their shape.[59,60] Nowadays it is possible to synthesize QDs with a wide variety of well-defined shapes, including spheres, stars, rods, rings, and tetrapods.[48,61–67] These anisotropic shapes can be obtained by oriented attachment,[61,68] kinetically induced anisotropic growth,[48,69] or seed-mediated solution-liquid-solid growth.[70] Moreover, the QD shape can also be influenced by additional components in the growth solution, like nanotubes or gold nanoparticles.[71–74] Anisotropic growth by oriented attachment was first described for TiO2 nanocrystals,[68,75,76] and is explained as follows: the nanocrystals continuously rotate until they find an identical crystal facet followed by spontaneous coalescence of two or more nanocrystals in order to reduce surface energy by elimination of high energy crystal facets.[77,78] The most common shapes

11

2 Background formed through oriented attachment are short nanorods and long nanowires,[61,79–81] but recently also more complex structures like honeycomb superlattices have been achieved.[82] Kinetically induced anisotropic growth is mainly influenced by the surfactants, however, the heating regime, or the monomer concentration also exerts an effect.[83,84] While spherical QDs form in the thermodynamic growth regime, the kinetic growth regime, which is achieved by a high growth rate, is necessary to form anisotropic QDs. At a low growth rate, all facets grow with an equal growth rate, due to only small differences in surface energies of low- and high-energy facets. Therefore, spherical QDs are formed in the thermodynamic growth regime. However, the growth rate depends exponentially on the surface energy, which means that for high growth rates high-energy facets grow faster than low-energy facets. This leads to an elimination of the fast growing facets, resulting in cubic QDs terminated solely by the slower growing facets.[48] If an organic surfactant adheres to a crystal facet it influences its growth process. The growth kinetics of different facets can be tuned using surfactants or surfactant mixtures with different binding energies, which allows for a precise tailoring of the final QD shape.[48] If the seed QD exhibits a polymorph crystal structure, which means that two or more crystal structures coexist in one nanocrystals, even branched nanocrystals (e.g., rod- or tetrapod shaped QDs) can be obtained.[69]

2.1.2 Optical and Electrical Properties The quantum size effect becomes obvious in the absorption spectra of QDs, where the position of the absorption peak is blueshifted with a decrease of the QD diameter (i.e., with an increase in the band gap) (Figure 2.5). Absorption spectra of monodisperse QDs not only show the first allowed transition, but also higher allowed transitions. Moreover, each sample of QDs exhibits a size dispersion of individual dots, causing a broadening of the absorption peaks. The narrower the size distribution of the QD sample, the more pronounced are the higher energy transition peaks and the smaller is the full-width half maximum (FWHM) of the first absorption peak. The photoluminescence (PL) peak, which results from a radiative recombination of an electron and a hole, is composed of band-edge and trap state emission. Some of the electrons generated by light absorption, first relax non-radiatively to energetically lower levels before they perform a radiative relaxation to the ground state. Thus, the wavelength of the band-edge emission maximum is lower in energy than the wavelength of the first absorption maximum. This redshift of the PL peak compared to the first absorption peak is called Stokes

12

2.1 Colloidal Semiconductor Quantum Dots shift. Trap state emission is even further redshifted than band-edge emission, as trap states are energetically located in the band gap.[85–89] There are many different sources that cause deep and shallow trap states in QDs. For example, the QD surface exhibits many dangling bonds due to the abrupt end of the crystal lattice. These dangling bonds act as trap states where charge carriers can become highly confined.[90] Moreover, additional acceptor or donor levels, which are generated by defects in the crystal lattice, can also trap charge carriers. The defects can be created during QD growth or after QD growth through surface oxidation.[91] Depending on the QD application, charge carrier trapping by surface or mid-gap trap states can be disadvantageous (e.g., traps are non-radiative recombination sites and therefore decrease light-emitting diode (LED) efficiencies[16] ) or beneficial (e.g., trap states generate photoconductive gain in photodetectors[92] or white light emission in LEDs[89,93] ).

(b)

increasing growth time

400

500

600

Wavelength (nm)

increasing growth time

Emission (a.u.)

(a)

Absorbance (a.u)

In summary, a high quality QD sample, defined by few trap states, good confinement of the exciton and a narrow size-distribution, can already be identified by its absorption and its PL spectrum. Such high quality QD samples are characterized by well-pronounced higher order absorption peaks, narrow absorption and PL band-edge emission peaks, no detectable trap state emission, and a small Stokes shift between the PL band-edge emission peak and the excitonic absorption peak.[12]

700

400

500

600

700

Wavelength (nm)

Figure 2.5: (a) Absorption and (b) emission spectra of a series of different-sized CdTe QDs. With increasing growth time, the QD size increases, leading to a decrease in the band gap and thus a redshift in absorption and emission.

In QD films, which are composed of different sized dots, energy transfer from smaller QDs to larger QDs is possible through dipole-dipole coupling. This F¨orster resonance energy transfer (FRET) causes a red-sift of film PL compared to solution PL and only happens if donor and acceptor are less than 10 nm away from each other. Therefore FRET is only observable in close-packed QD films and not in QD solutions.[22,94]

13

2 Background The extremely bright PL[95] of most QDs in solution is quantified by the photoluminescence quantum yield (PL QY). The PL QY is defined as the ratio of the number of emitted photons to the number of absorbed photons. The PL QY of typical cadmium chalcogenide QDs synthesized and measured in solution is higher than 50 %, while it decreases to around 5-10 % in cadmium chalcogenide thin-films.[90] This decrease in PL QY can be explained by the larger effect of non-emitting QDs in thin-films than in solutions. In QD thin-films only one non-emitting dark excitonic state is able to quench the emission of a few highly luminescent neighboring QDs, due to FRET processes.[90] Moreover, in QD thinfilms, unlike to QD solutions, surface trap states cannot be passivated by an excess of ligands.[90] However, the PL QY of QDs is not only influenced by trap states, but also the surfactants,[96] a possible surfactant exchange,[88,97] and the solvents used during the washing process[57] can either increase[96,98,99] or decrease[57] the PL intensity. Additionally, the surface stoichiometry greatly influences the PL QY of QDs.[100,101] Despite of all difficulties and quenching effects, QDs with their easily tunable optical properties, their spectrally narrow and efficient photoluminescence, and their good photostability are optimal candidates for light emitting and detecting materials in optoelectronic applications.

Linear and non-linear exciton dynamics - Excitons, multiple excitons, and trions. First, the linear exciton dynamic range will be considered. If exciton energies are low and photons with energies equal or higher than the band gap hit a QD, only one exciton per photon is generated on average. An exciton is a neutral electron hole-pair, which is bound by Coulombic attractions (Figure 2.6(a)). Excitons can undergo different relaxation pathways.[12] If the excitation energy is much higher than the band gap energy, electrons and holes exhibit an excess kinetic energy (so called “hot” charge carriers) which can be lost by rapid intraband electron or hole thermalization, to the CB and VB edges.[102] Intraband relaxation is a very fast process and happens on the timescale of several tens to hundreds of femtoseconds. Electrons and/or holes from the VB and CB edges may be trapped and recombine after a certain trapping time or recombine instantly without trapping. Recombination can either happen radiatively, resulting in the emission of a photon or non-radiatively by releasing heat. High quality QDs with almost no trap states exhibit a high PL QY. This high PL QY results from a mainly radiative recombination with lifetimes of a few to a few tens of nanoseconds. If charge carriers are trapped, trapping (a few hundred femtoseconds to a few tens of picoseconds) is faster than the bandedge radiative recombination

14

2.1 Colloidal Semiconductor Quantum Dots and therefore decreases the PL QY. Once a charge carrier is trapped, it can either recombine or become trapped in even deeper trap states, with lifetimes of hundreds of picoseconds to microseconds.[12,103] If excitation energies are increased, exciton dynamics change from linear to non-linear. In the non-linear range one photon can generate more than one exciton, e.g., a biexciton (Figure 2.6(b)). In 2004 Schaller et al. reported for the first time multi-exciton generation (MEG) in QDs.[104] MEG identifies the process, where one photon can generate more than one exciton. In order to achieve MEG the photon energy has to be at least two times the band gap energy. Therefore, small band gap lead chalcogenide QDs are ideal materials to observe MEG.[105–108] Non-linear exciton processes are for example Auger recombination and exciton-exciton annihilation. During Auger recombination the energy released by exciton recombination is transferred to a third charge carrier (either an electron or a hole), which is then excited to a higher energy state and subsequently thermalizes to the first excited state.[90,109,110] Exciton-exciton annihilation[102] is suppressed if charge carrier trapping occurs faster than exciton-exciton annihilation. Only when all trap states are saturated, exciton-exciton annihilation starts to happen, which is why the threshold for exciton-exciton annihilation indicates the trap state concentration in QDs. Charge carrier trapping at low pump intensities and Auger recombination and excitonexciton annihilation at high pump intensities are fast, non-radiative processes, which all decrease the overall charge carrier lifetime and PL QY of QDs.[12,109] Neutral QDs can be turned into trions, which are charged three-particle bound states of a neutral excitonic electron-hole pair plus an extra unpaired hole (positive trion) or electron (negative trion) (Figure 2.6(c) and (d)). Trions can be generated by electrochemical charge injection, by certain chemical treatments (i.e., reducing or oxidizing species), by photoexcitation, or by excitation with energetic electrons, and may decay via Auger recombination.[111] Jha et al.[112] reported that the Auger lifetime of trions should be four times longer than the Auger lifetime of biexcitons,[111,113] assuming an equal effective mass for electrons and holes. Although trion Auger lifetimes are much shorter than exciton radiative lifetimes, they can also recombine radiatively contributing to the band-edge PL.[114,115]

2.1.3 Quantum Dot Solids In all kind of applications, QDs will never appear as individual particles, they rather exist as QD films. Therefore, it is very important to be able to fabricate homogeneous films, to fully understand and to controllably engineer QD coupling

15

2 Background

Figure 2.6: Schematic illustration of excitons, biexcitons and trions. (a) An exciton is a neutral electron-hole pair, while (b) biexcitons are two electron-hole pairs, still neutral in their overall charge. Trions are charged three-particle bound states of a neutral excitonic electron-hole pair (c) plus an extra unpaired hole (positive trion) or (d) electron (negative trion).

and charge transport.[116] QD films have two major drawbacks that have to be overcome in order to build efficient devices. On the one hand, they exhibit poor charge carrier mobilities due to a low electronic coupling within the QD solid. On the other hand, QDs normally have an imperfect passivation and structural defects causing a high density of surface- and midgap-trap states that lead to non-radiative losses.[117] In order to use QDs in commercial applications it is essential to improve coupling, reduce trap states, and to optimize thin-film morphology.[116] Solution-based material processing of QDs - Thin-film formation. QDs can be dispersed in volatile solvents, therefore it is relatively simple to process the material and to form QD thin-films on various substrates. These thin-films are also known as QD solids. QDs with a narrow size distribution and a spherical shape will form hexagonal close-packed domains (10-100 nm), enabling a good interdot coupling.[15,56,118] Different methods are used for film formation, like spin coating,[1] dip coating,[2,3] drop-casting,[4,5] inkjet printing,[6–8] or spray coating.[9,10] While drop-casting results in relatively inhomogeneous films, spin- and dip-coating yield homogeneous QD solids with a highly controllable film thickness. Improve charge carrier transport. A basic requirement for interdot coupling and thus an efficient charge carrier transport is a homogeneous distribution of QD energy levels. Such a homogeneous energy level distribution is present in QD samples with a narrow size distribution.[118] Nevertheless, QD films prepared with equally sized QDs are also insulating for electrical transport, as QDs are covered with long-chain organic molecules that provide a steric stabilization of the QDs in solution but at the same time prevent strong coupling and charge

16

2.1 Colloidal Semiconductor Quantum Dots transport between the QDs.[56,119] Due to the vanishingly low interdot coupling, the generated excitons will stay confined within a single QD and will not be separated into free electrons and holes.[120] In order to overcome this limitation, there are different ways to improve the coupling in QD solids. The most prominent approach to improve coupling is to perform a ligand exchange, where the initially long insulating ligands are replaced with short ligands in order to reduce the interparticle spacing.[121,122] Thereby it is important to keep the discrete and size-dependent features of the QDs. Suitable short ligands are small organic molecules like thiols (e.g., ethandithiol (EDT), benzenedithiol (BDT))[123–135] amines (e.g., hydrazine, butylamine, methylamine, hexylamine),[3,4,46,116,119,124,125,136–138] pyridine,[139] organic acids (e.g., 3mercaptopropionic acid (MPA)),[126,140,141] or small inorganic capping molecules like metal chalcogenide complexes,[142–146] oxo- and polyoxometallates,[147,148] lead halide perovskites,[149] metal halide complexes,[149] and metal free ions (e.g., halides, thiocyanate, sulfide).[118,150–154] The inorganic molecules, which were first used in 2009, enable the formation of all-inorganic QD solids. They strongly enhance the electronic transport by not only forming crack-free (little weight loss of ∼ 3.8 %) and completely inorganic films but also through a favorable highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) engineering, leading to very high charge carrier mobilities.[118,142,143,145,150,155] Ligands can be exchanged via two strategies: pre-film-deposition via a solution phase ligand exchange[4,46,142] or post-deposition in a solid state ligand exchange.[3] With solution phase ligand-exchanged QDs it is possible to form crack-free and dense QD solids via a single processing step. However, the short ligands introduced through the ligand exchange often decrease the colloidal stability of the QDs, therefore these dots tend to aggregate already before film formation. A solid state ligand exchange involves two steps: first, QD film formation and second, treatment of the QD film with a solution containing the exchanging ligand. The volume loss caused by exchanging long ligands with short ligands results in QD films with many cracks. These cracks may short-circuit the device when an electrode is evaporated on the film and therefore cracks have to be avoided. In order to infill the cracks and achieve a continuous film, multiple layers produced via a layer-by-layer (LBL) process are necessary. In such a process the QD deposition and the following ligand exchange are repeated a few times.[116] The thickness of the film can be controlled either via the number of layers, the spin- or dip-coating speed, or the concentration of the QDs in solution.

17

2 Background A ligand exchange not only improves interdot coupling, it also increases the number of surface dangling bonds, which are responsible for trap states. Therefore a ligand exchange decreases PL efficiency and might also decrease charge carrier mobility by charge carrier trapping.[56] Moreover, short ligands render QDs more prone to oxidation compared to QDs with long ligands.[15] Thus, a good protection from oxidation and a controlled surface trap passivation are very important to achieve QD solids with high charge carrier mobilities.[118] A common way to reduce trap states and efficiently passivate the QD surface is to introduce halide anions.[150,156,157] Alternatively, hydrazine saturates surface dangling bonds due to its lone pair of electrons and additionally decreases the oxidation possibility of the QD surface.[15] Other strategies to improve charge carrier transport in QD solids include film annealing, chemical doping, QD shape optimization, QD film optimization, or QD hybrid formation. The annealing of a QD solid leads to a loss of ligands. This ligand loss reduces the interparticle spacing and increases the tunneling rate between the QDs but at the same time broadens and redshifts the excitonic peak due to a fusion of neighboring QDs. This QD fusion causes a partial loss of quantum confinement.[158–160] Law et al. discovered that PbSe QDs are thermally not stable enough to achieve temperatures that are necessary for oleate pyrolysis.[116] Therefore, in order to lose oleic acid, a ligand exchange prior to annealing is anyway necessary.[161] Moreover, an annealing process is not compatible with most polymer substrates, which are necessary for flexible electronics. The improvement of charge transport by chemical doping is based on the increased mobile carrier concentration,[15] for example by adding additional charge carriers to the QD film.[118,162,163] As reported by Choi et al., the combination of a ligand exchange with a doping of the QD film results in an extremely high charge carrier transport.[118] However, due to some challenges like self-purification of QDs, a controlled doping is difficult to achieve.[164] Shape optimization aims to synthesize cubes, wires, sheets, or hybrids in order to reduce the number of hopping events between QDs and thus improves charge carrier transport.[15] QD film optimization is based on a selective ligand displacement leading to a self-assembly of QDs through epitaxial necking of similar crystal facets, forming confined-but-connected superstructures.[165,166] A completely different strategy to improve charge carrier transport is to make use of hybrid structures, where QDs only act as light absorbing material, while a second material is responsible for the charge carrier transport (see Section 2.4 and Chapter 4).

18

2.1 Colloidal Semiconductor Quantum Dots Charge carrier transport in QD films. Charge carrier transport in QD films is a hugely investigated topic[136,167–172] and nowadays it is believed that the charge carrier transport changes from hopping between localized states in weakly coupled QD films (Figure 2.7(a))[133,169,173,174] to coherent band-like transport in strongly coupled QD films (Figure 2.7(b)).[120,175,176] The hopping probability of charges between QDs depends on the interparticle distance and the interparticle material. The interparticle distance is associated with the tunnel barrier width, while the interparticle material represents the tunnel barrier height. Therefore, by decreasing the ligand length (assuming the same anchoring group), the tunneling barrier is reduced and the charge carrier mobility increases exponentially. Another way to increase the hopping probability is through a reduction of the tunneling barrier height by replacing the initially saturated organic ligands with conjugated organic ligands or inorganic ligands.[2,125,172,177,178] However, hopping of charge carriers not only happens between neighboring QDs, it also occurs over longer distances, then called variable range hopping.[179] If charge transport happens via hopping, the charge carrier mobility increases with QD diameter due to two different effects. First, fewer hopping events are required to transport charge carriers through a certain distance covered with large QDs compared to the same distance covered with small QDs. Second, a decreasing band gap comes along with a decreasing depth of trap states, retaining the charge carriers less confined.[2] Charge carrier hopping occurs non-adiabatically between QDs. This means that a small activation energy is always necessary to achieve charge carrier hopping. Even if the activation energy decreases for high excitation densities, it will never disappear.[120,177] Recently, a charge carrier transport with a negative temperature coefficient strongly indicated a band-like transport in QD solids. Band-like transport happens through extended electronic states, which are no longer localized on single QDs. Additionally, a combination of Hall-effect measurements and field-effect transistor characterizations were used to corroborate band-like transport in QD solids.[120,177,180] It has to be noted that Hall mobilities are usually strongly suppressed in the hopping transport regime.[177,181] All high-mobility QD devices (µe in CdSe QDs ∼ 38 cm2 V−1 s−1[155] ) were fabricated with inorganically capped QDs,[118,142,145,151,152,155] which clearly indicates that improving QD surface chemistry is a major tool to improve charge transport. Band-like charge transport in strongly coupled QD films approaches the mobilities of crystalline bulk semiconductors, while still keeping the unique quantum confined properties of QDs.[118] That means that the absorption bands of the QDs are still present in the absorption spectra of QD solids, they are only slightly redshifted and broadened due to electron delocalization.[118] The realization

19

2 Background of high-mobility QD solids opens up new possibilities for commercial QD device applications. (a)

(b)

„separated“ quantum dots

„connected“ quantum dots

Figure 2.7: Schematic illustration of charge carrier transport (a) via hopping between localized states in weakly coupled QD thin-films and (b) via coherent band-like transport in strongly coupled QD thin-films. Figure adapted from Choi et al.[118]

2.1.4 Lead Chalcogenide Quantum Dots Lead chalcogenide QDs (PbX with X = S, Se, or Te) are IV-VI semiconductors based on lead cations and chalcogenide anions. In 2001 Murray et al.[182] performed the first hot-injection synthesis for colloidal lead chalcogenide QDs. The PbSe QDs were synthesized with a size range between 3.5 and 15 nm by the injection of lead oleate and tri-n-octylphosphine (TOPSe) in hot diphenylether. A subsequent sizeselective precipitation narrowed the size dispersion of the samples.[182] The work by Murray et al. paved the way for the first hot-injection synthesis of PbS QDs in 2003 by Hines et al.[183] High-quality PbS QDs with a PL QY of around 20 %, a FWHM of the PL peak of 100 meV, and a size distribution of 10-15 % without post-synthesis size-selective precipitation was achieved with a synthesis based on lead(II) oxide (lead precursor), bis(trimethylsilyl)sulfide (sulfur precursor), oleic acid (surfactant), and octadecene (solvent).[183] These two syntheses were followed by numerous slightly modified syntheses to produce monodisperse PbS,[27,184] PbSe,[173,185,186] and PbTe QDs.[187–189] In the present thesis, only PbS and PbSe QDs were used, therefore all of the following properties are only reported for those two materials, neglecting PbTe. Bulk PbS and PbSe are narrow band gap semiconductors with direct band gaps of 0.41 eV[190,191] and 0.28 eV,[30,186,192] thus absorbing and emitting light in the near-infrared (NIR) region. They exhibit large excitonic Bohr radii (PbS 18 nm,[191] PbSe 46 nm[192] ), which makes it relatively easy to access the strong quantum confinement regime, also for large QDs.[14,193] Due to their nearly identical and

20

2.1 Colloidal Semiconductor Quantum Dots large Bohr radii for electrons and holes (23 nm for PbSe[14] and 10 nm for PbS[28] ), PbS and PbSe QDs are able to simultaneously transport both charge carriers. This property renders PbX QDs an ambipolar material. Lead chalcogenide QDs can be synthesized with a wide size range between 2-10 nm,[194] corresponding to a tunable absorption, which ranges from around 750 to 2400 nm.[17] PbX QDs exhibit size dependent PL QYs, which reach 80 % for small QDs[27] and decrease to less than 3 % for large QDs.[194] PbX QDs possess a centrosymmetric rock salt crystal structure, where the surface of the PbX crystals has six nonpolar {100} facets and eight polar {111} facets. The {111} facets are either terminated by lead cations or chalcogenide anions. Depending on the distribution of chalcogenide rich or lead rich facets, statistically 89 % of lead chalcogenide QDs have a dipole moment, most common in the h001i direction (Figure 2.8).[61,195,196] Moreels et al.[191,197] showed that PbX QDs exhibit a nonstoichiometric Pb/X ratio, which results from an excess of lead atoms at the surface of the QDs, while the core is quasi-stoichiometric.[192,198,199] It is important to carefully tailor the stoichiometry of PbX QDs, because it greatly influences the electrical properties.[200]

Figure 2.8: Schematic illustration of the dipole moment in centrosymmetric lead chalcogenide QDs. PbX QDs exhibit eight polar {111} facets, either terminated by lead cations (“green”) or chalcogenide anions (“gray”). Depending of the distribution of the eight polar {111} facets, statistically 89 % of lead chalcogenide QDs have a dipole moment, most common in the h001i direction (red arrow). Figure adapted from Cho et al.[61]

PbX QDs are sensitive to air and humidity and tend to oxidize quickly. A short exposure of a PbX QD thin-film to air results in an adsorption of O2 and H2 O. These two compounds p-dope the QD film, leading to a PL decrease and an increase in film conductivity. However, after a longer exposure to oxygen an oxide shell is created and PbX QDs oxidization is irreversible. The passivating oxide shell drastically decreases the film conductivity, while it induces a recovery of the PL intensity. Surface oxidation can be monitored by absorption spectroscopy, because the reduction of the effective diameter of the PbX core through oxidation causes a blueshift of all absorption bands.[178,201–203] Smaller QDs are less prone to oxidation than larger QDs, because the higher lead concentration at the surface of small QDs

21

2 Background compared to large ones results in a denser packed ligand shell and thus a better surface passivation.[128] A good surface passivation, e.g., with halide ions, improves the air stability of PbX QDs of all sizes and partly provides oxidation protection and trap state filling.[17,130,150,157] Halide passivation can happen post-synthesis by adding a halide source (e.g., tetrabutylammonium iodide, CdCl2 )[156,204,205] or pre-synthesis by using PbCl2 as lead precursor.[184,206,207] However, to exclude oxidation and loss of ambipolar behavior, PbX QDs should be processed and measured in an inert atmosphere. The synthetic procedures play a crucial role in the quality, size, and shape of PbX QDs. A high injection temperature results in larger QDs compared to a low injection temperature.[17] Moreover some ligands (e.g., oleic acid) promote a spherical growth while others (e.g., oleylamine) favor the formation of cubic QDs through a higher growth rate in the h111i direction than in the h100i direction.[208,209] Additionally, anisotropic PbX QDs can be grown via oriented attachment,[61] which is driven by a combination of dipole-dipole interactions and the reduction of surface energies.[61,195,210,211]

2.2 Carbon Nanotubes Carbon nanotubes (CNTs) are one-dimensional carbon allotropes, that were first observed in 1991 by S. Iijima from deposits on graphite cathodes after an arcdischarge process, which initially intended to produce the zero-dimensional carbon allotrope fullerene.[212] The CNT structure can be described by a rolled graphene sheet, that consists of sp2 -hybridized carbon atoms, arranged in a honeycomb lattice.[213,214] The honeycomb lattice is formed by covalent σ-bonds in the xy-plane between three carbon atoms, which are arranged in a 120◦ angle to each other. Moreover, pz orbitals form delocalized π-bonds perpendicular to the wrapped sheet. These π-bonds are responsible for the van-der-Waals interactions between single CNTs and the electrical conductivity in CNTs.[213] CNTs occur as single-walled carbon nanotubes (SWNTs) and multi-walled carbon nanotubes (MWNTs). In SWNTs a single graphite sheet (i.e., graphene) is rolled into a cylindrical tube, whereas MWNTs are a stack of single-walled carbon nanotubes held together by weak van-der-Waals forces.[215] The first CNTs reported by Iijima et al. were MWNTs,[212] followed by a gas-phase growth of SWNTs two years later.[216] The structure of CNTs is determined by a pair of integers (n, m), which defines the chiral vector Ch = na1 + ma2 ≡ (n, m), 22

(2.4)

2.2 Carbon Nanotubes where a1 and a2 are the unit vectors of graphite.[217] Based on the chiral vector the diameter can be defined as d=

a√ 2 |Ch | = n + nm + m2 , π π

(2.5)

√ ˚).[217] The way of rolling the graphene where a is the lattice constant (1.42 x 3 A sheet results in three different carbon nanotube types, namely zigzag (m = 0), armchair (n = m) and chiral (n 6= m) (Figure 2.9).[215,217] If (n−m)/3 is a nonzero integer, CNTs exhibit metallic properties, whereas in all other cases they are semiconducting. Overall, two thirds of the CNTs are semiconducting, while one third are metallic.[215,217–221] In this work, exclusively SWNTs were used. SWNTs can have a diameter between 0.4 and 3 nm[215] and a length up to a few millimeter,[222] resulting in high aspect ratios of 105 -106 .

a2

unit vector

a1

A

(n,n)

armchair

B

unit cell

0

q

Ch

chiral vector

(n,0)

zig zag

chiral angle

Figure 2.9: Schematic illustration of a graphene lattice, consisting of sp2 -hybridized carbon atoms, arranged in a honeycomb lattice. CNTs are formed by rolling the graphene sheet along the chiral vector Ch with a certain chiral angle θ. The chiral vector is defined by the two integers (n, m), determining armchair (θ = 30), zig-zag (θ = 0), or chiral CNTs. Additionally depicted are the unit vectors a1 and a2 and the hexagonal unit cell, containing two carbon atoms (A and B). Figure adapted from Jakubka.[223]

2.2.1 Fabrication and Processing Carbon nanotube fabrication. CNTs are produced via carbon arc discharge,[212,224] laser ablation of carbon,[225] or chemical vapor deposition (CVD).[226–228] The resulting CNTs usually occur in bundles of parallel tubes, attached to each other via van-der-Waals forces.[225] The so-produced CNTs differ in their diameter, length, quality, and purity. So far all synthetic routes lead to a mixture of semiconducting and metallic SWNTs.[215] In the carbon arc discharge method CNTs are generated

23

2 Background between graphite rods, while for the laser ablation technique a laser beam with a high intensity is focused on graphite.[229] In this work, only CNTs fabricated by CVD processes, namely powders of commercially available HiPCO[230] and CoMoCAT[231] SWNTs, were used. In principle, a CVD process to produce CNTs is based on a heterogeneous reaction, using a carbon source and a metal catalyst. In detail, a hydrocarbon gas flows around the catalyst material (e.g., Fe, Co, Ni) and the CNTs subsequently grow from the catalyst by thermal disproportionation of the hydrocarbon gas. HiPCO stands for high-pressure carbon monoxide conversion, a large-scale gas-phase CVD process, where the primary carbon source is carbon monoxide and the metal catalyst is an iron cluster. The iron cluster is generated in situ by thermal decomposition of iron pentacarbonyl at temperatures around 900-1100 ◦ C. The high pressure of 30-50 atm leads to a high product yield of SWNTs of up to 97 %.[230,232] CoMoCAT nanotubes are produced from CO with a mixed catalyst from cobalt nitrate and ammonium heptamolybdate on a silica support. The continuous flowing carbon source decomposes at high temperatures and the Co-Mo catalyst acts as nucleation center to start the CNT growth.[231] CNTs fabricated by carbon arc discharge or laser ablation usually exhibit less structural defects than CNTs grown by CVD,[233] but contain more impurities like carbon nanoparticles, amorphous carbon, fullerenes, and metal particles. These impurities can partly be removed by selective oxidation, acid treatment, centrifugation, or filtration.[229]

Carbon nanotube processing. The CNTs obtained by HiPCO or CoMoCAT processes are in powder form, where the tubes are normally bundled and need to be separated.[225] Separation is done by a liquid exfoliation process using ultrasound treatment to break up the van-der-Waals forces between the single nanotubes followed by ultracentrifugation to separate debundled CNTs from bundled CNTs. However, ultrasound treatment breaks some CNTs, which decreases their length and thus their aspect ratio.[213,234] CNTs are barley soluble in solvents, therefore they need stabilizers which adsorb non-covalently to their surface in order to form stable dispersions. Depending on the stabilizer, CNTs can be exfoliated in organic or aqueous media. Tenside detergents like sodium dodecyl sulfate (SDS)[235–238] and sodium dodecylbenzene sulfonate (SDBS),[239] deoxyribonucleic acid (DNA),[240] or some conjugated polymers[241] help to stabilize CNTs in water, while some other conjugated polymers[242–244] form stable CNT-organic solvent dispersions. Moreover, Ausman et al. reported that CNTs also exhibit a moderate solubility in the organic solvents n-methyl-2-pyrrolidone (NMP) and dimethylformamide (DMF).[245] Since

24

2.2 Carbon Nanotubes a growth of neither only semiconducting or metallic CNTs nor a selective growth of single chirality CNTs was successful up to now, a post-synthesis step is always necessary to separate semiconducting from metallic CNTs or to obtain a single chirality. The controlled access to certain chiralities of CNTs is important in order to make them applicable for (opto)electronic devices.[246] Some separation methods include dielectrophoresis, ultracentrifugation, chromatography, selective destruction, or selective interaction with certain molecules.[244,247–253] However, methods to separate semiconducting from metallic CNTs are costly and timeconsuming.

2.2.2 Optical and Electrical Properties Optical techniques, like absorption, PL and Raman spectroscopy are ideal for the characterization of CNTs, because they are non-destructive and at the same time provide a lot of information about quality, quantity, and optical and electrical properties of the CNTs.[254] Absorption and PL spectroscopy. CNTs are a one-dimensional material, where electrons and holes are confined in two-dimensions. This one-dimensionality is responsible for the well-defined sharp peaks (i.e., van-Hove singularities) in the density of states distribution of CNTs (see Figure 2.2).[214,255,256] Transitions between the sub-bands lead to certain features in the absorption spectra of CNTs.[235,257] The energetically lowest feature is the metallic band M11 arising between 400600 nm. The first semiconducting band S11 appears at 800-1600 nm, generated by transitions from the highest occupied state to the lowest unoccupied state, slightly overlapping with the second semiconducting feature S22 at 550-900 nm, which is caused by higher order transitions.[235,257] The better the nanotubes are separated, the narrower are the absorption peaks within the particular bands, while bundles of CNTs cause a merger of the absorption peaks to broad absorption bands.[258] The tube diameter is inversely proportional to the band gap, leading to a redshift of the absorption with increasing CNT diameter. Therefore, absorption spectroscopy is an important tool to identify individual SWNT chiralities.[222,257] All optically allowed transitions for many different SWNTs are summarized in the Kataura plot, which depicts the energy versus the CNT diameter.[256,257] Moreover, the ratio between semiconducting and metallic CNTs can roughly be deduced from the intensity ratio of S to M peaks.[257] Semiconducting SWNTs exhibit a direct band gap, thus PL should be observable. However, bundles of

25

2 Background SWNTs are often composed of a mixture of semiconducting and metallic SWNTs, where the metallic SWNTs quench the emission of the semiconducting tubes. Only when a complete debundeling of SWNTs is achieved, PL of semiconducting SWNTs can be detected.[235,255,256] Raman spectroscopy. Raman spectroscopy is a technique that relies on the inelastic scattering of light. In order to see bands in a Raman spectrum, the investigated molecule must be able to change its polarizability.[259] Raman bands are particular pronounced in resonance Raman spectroscopy, where the excitation energy matches the energy of the optical allowed transitions.[256] SWNTs exhibit four characteristic Raman bands. The radial breathing mode (RBM) is generated by atomic vibrations in the radial direction and appears between 120 cm−1 and 250 cm−1 (for SWNTs between 1-2 nm). The exact position depends on the diameter of the nanotube, therefore the RBM band can be used to determine the CNT diameter.[254,256,260] The disorder-induced D-line around 1340 cm−1 indicates defects in the carbon lattice and increases in intensity with an increasing defect density in the nanotubes. The G’-line at around 2600 cm−1 is a second order harmonic of the D-band. The fourth CNT specific Raman band is a graphite related optical mode, which is called G-band and appears at 1550-1600 cm−1 .[256] In contrast to graphite, where the G-band is a single band with one peak at 1582 cm−1 , the G-band in CNTs is composed of two features. One feature is caused by vibrations along the tube surface (G+ ), whereas the second feature is generated by vibrations orthogonal to the tube surface (G− ). The width of the G− feature increases with an increasing ratio of metallic CNTs and thus indicates if the CNT sample contains mostly semiconducting or metallic CNTs. Moreover, the G− band decreases with decreasing CNT diameter, while the G+ band is unaffected by the diameter.[256,261] Electrical properties. The electronic transport along the axis of SWNTs occurs almost ballistcally, i.e., without scattering by defects or lattice vibrations. Ballistic transport is possible if the mean free path is larger than the length of the CNT. Therefore, ballistic transport happens in short defect free tubes, whereas in longer tubes many collisions may take place, leading to a diffusion limited process.[214,221,262] Charge carrier transport in CNTs occurs by holes and electrons in inert atmosphere and is reduced to predominantly hole transport in ambient air due to oxygen related doping effects. The charge carriers propagate along the tube axis through the carbon π-orbitals,[222] resulting in very high carrier mobilities of

26

2.3 Layered Materials around 100.000 cm2 V−1 s−1 and high on/off ratios of 106 for individual CNTs.[263] These carrier mobilities are as much as 1000 times higher than carrier mobilities in bulk silicon.[214] However, there are some limitations in making use of these properties in actual devices. Depending on the nanotube type, there is a large contact resistance to metal electrodes due to Schottky barriers and parasitic resistances.[264] Moreover, it is very costly and time-consuming to separate semiconducting from metallic nanotubes.

2.3 Layered Materials In 2004 Novoselov et al.[265] published the first experimental evidence of graphene, a two-dimensional layer of carbon atoms exfoliated from graphite (Figure 2.10). Theoretical research on two-dimensional graphite has been performed for almost seventy years.[266,267] However, up to 2004 it was assumed that producing freestanding graphene is impossible, because it is unstable and will instantly form curved structures.[265] Novoselov et al. succeeded in producing graphene by mechanical exfoliation of a highly oriented pyrolytic graphite with a Scotch tape. The experimental discovery of graphene caused enormous research efforts not only about graphene itself, but also about other two-dimensional materials like transition metal dichalcogenides (TMDs) (e.g., MoS2 , WS2 ) (Figure 2.10) and boron nitride.[268] The breakthrough results by Novoselov et al. paved the way for an intense research about the properties of all kind of two-dimensional materials. It soon became clear that two-dimensional layers exhibit very interesting optical and electrical properties that differ from bulk material properties due to confinement effects in one dimension. The DOS is proportional to the dimensionality d according to: DOS(E) ∼ E

(d−2)/2

, with d = 1, 2, 3,

(2.6)

where E is the energy. Therefore, the DOS of two-dimensional materials describes a step-like function, independent of the energy E (see Figure 2.2).[38]

2.3.1 Graphene Graphene is the two-dimensional form of the three-dimensional carbon allotrope graphite. Graphite is a layered material with a stacked structure, where the adjacent two-dimensional graphene layers are held together by weak van-derWaals forces, while there are strong covalent interactions within each graphene 27

2 Background (a)

(b)

Top-view

Top-view S Mo/W

C

Side-view

2 layer

Side-view 0.335 nm

2 layer

0.65 nm

Figure 2.10: Top-view and side-view schematics of the structure of (a) few-layered graphene and (b) MoS2 (WS2 ). The adjacent two-dimensional layers are held together by weak van-derWaals forces, with a layer spacing of around 0.335 nm for few-layered graphene and around 0.65 nm for MoS2 (WS2 ).

layer. Therefore, graphite can easily be cleaved along the layer surface, leading to graphene.[269] Graphene is composed of a single atom thick two-dimensional layer of sp2 -hybridized carbon atoms arranged in a honeycomb lattice.[222,265] Each carbon atom forms three σ-bonds to neighboring carbon atoms and one π-bond with the p orbital orientated perpendicular to the carbon layer, resulting in a long-range delocalized π-conjugation. This delocalized π-bond is responsible for charge carrier conductivity along the graphene sheet.[269] Graphene is regarded as a two-dimensional material up to ten monolayers, beyond this number it is approaching the three-dimensional limit.[270,271] Graphene properties. Graphene is a zero-gap semiconductor with many exceptional properties like a large specific surface area,[272] an optical transmittance of 97.7 %,[273] a thermal conductivity of 5300 W/mK,[274] and a high mechanical strength (Young’s modulus of 1 TPa).[275] Moreover, its charge carriers behave as massless Dirac fermions,[276,277] this is why graphene shows extremely high carrier mobilities, reaching micrometer scale ballistic transport.[265,278,279] Suspended mechanically exfoliated graphene has charge carrier mobilities of over 200.000 cm2 V−1 s−1 .[280,281] Although these extremely high carrier mobilities render graphene interesting for electronic devices, its low on/off ratio of around 2-10, due to its zero band gap, limits the usability in devices that need to have clear on and off states.[222] Single- and few-layer fabrication. There are many techniques to produce singleand few-layer graphene. Depending on the desired application of the graphene sheet, the best suitable technique has to be chosen carefully. The primary method to produce graphene was micromechanical cleavage of a highly oriented pyrolytic 28

2.3 Layered Materials graphite using Scotch tape. This method yields high quality graphene sheets, however the process is time-consuming and has low efficiencies.[265] In 2006 Berger et al. published a method to grow graphene epitaxially by thermal decomposition of a silicon carbide surface. With this technique high quality graphene is grown at an expense of high costs, high production temperatures, and a difficult transferability to other substrates.[282,283] Moreover, graphene can be fabricated by CVD growth via thermal decomposition of a carbon source on a metal catalyst (e.g., Ni, Cu). CVD leads to large single- or few-layer graphene sheets with few defects and impurities that can be transferred to many other substrates.[284–286] Another common method to obtain graphene is by liquid phase exfoliation of graphite or graphite oxide. Solution-based exfoliation techniques are cost-efficient and compatible with mass-scale production. The oxidation of graphite increases the layer spacing, which results in weaker interactions between the layers. Therefore, exfoliation by ultrasonic treatment or mechanical stirring is very efficient leading to single graphene oxide layers. These layers may be subsequently chemically or thermally reduced, in order to form reduced graphene oxide (rGO) layers.[287] Liquid phase exfoliation via graphene oxide generates high yields, however the final graphene layers show many defects which were induced during the oxidation step. These defects decrease the charge carrier mobility but at the same time also provide a good solubility of rGO in polar solvents.[288,289] The solution-based exfoliation using non-oxidized graphite can be realized with certain organic solvents like NMP, DMF, and o-dichlorobenzene (oDCB).[290–292] These solvents have a surface tension that promotes the increasing total area of graphite.[293] The soproduced graphene sheets show fewer defects than rGO, but the overall production efficiency is very low. Analogous to CNTs, graphite can also be exfoliated in an aqueous/surfactant mixture, e.g., water/SDBS.[294] Other methods to produce graphene include the bottom up organic synthesis from benzene rings[295] or electromechanical exfoliation.[296]

Raman spectroscopy. Raman spectroscopy provides a possibility to identify the quality and the number of graphene layers in a nondestructive way. It is possible to distinguish between single-layer, bilayer, and few-layered graphene (FLG) (≤ 5 layers). For five and more layers the Raman spectrum of graphene approaches the one of bulk graphite.[297] Raman spectra of graphene show three intense Raman features, the graphite related G-band at ∼ 1580 cm−1 , the disorderinduced D-band around 1350 cm−1 , and the 2D-band at around 2700 cm−1 (Figure 2.11).[297] The shape and position of the 2D-band and the intensity ratio of the

29

2 Background 2D-band to the G-band indicate the number of graphene layers. Graphene exhibits a single and sharp 2D-peak, however, with an increasing number of stacking layers, the 2D-band broadens and shifts to higher wavenumbers.[297] For single layer graphene the 2D-band can be fitted with a single Lorentzian function with a FWHM of ∼ 24 cm−1 , whereas the same band can be fitted with four Lorentzian functions (each having a FWHM ∼ 24 cm−1 ) for bilayer graphene.[297,298] Moreover, the G-band in graphite is much more intense than the 2D-band, therefore a large intensity of the 2D-band compared to the G-band indicates an approach of graphite to single-layer graphene.[297–299]

Intensity (a.u.)

2.0 1.5

2D G

1.0 0.5

D

0.0 2000

3000 -1

Raman shift (cm ) Figure 2.11: Raman spectrum of FLG, showing three intense Raman features: the graphite related G-band at ∼ 1580 cm−1 , the disorder-induced D-band around 1350 cm−1 , and the 2D-band around 2700 cm−1 .

2.3.2 Transition Metal Dichalcogenides Research interest in layered materials is already quite old[300] and first reports about single-layer MoS2 date back to 1986.[301] However, the interesting material group of layered compounds, including TMDs, regained popularity with the experimental discovery of graphene in 2004.[265] In 2005 Novoselov et al. extracted layers of two-dimensional TMDs by rubbing the layered crystals against another surface on which TMD single-layers were found afterwards.[268] Similar to graphite, TMDs are van-der-Waals bonded layered materials with a layer spacing of about 6-6.5 ˚ A (see [302–307] Figure 2.10). One charge neutral layer can be considered as three covalently bonded atomic sheets, a positively charged two-dimensional hexagonal lattice of metal atoms sandwiched between two hexagonal lattices of negatively charged chalcogenide atoms with a X-M-X stoichiometry (M = group IV element, e.g., Mo, W; X = S, Se, Te). Each metal atom forms six covalent bonds to chalcogenide atoms, whereas each chalcogenide atom covalently bonds to three metal atoms. The surface of each layer is terminated by a lone-pair of electrons of the chalcogenide

30

2.3 Layered Materials atoms, but due to the absence of dangling bonds the layers are unreactive to environmental species.[308,309] TMD properties. TMDs are semiconductors, with an indirect band gap in the three-dimensional bulk crystal (Eg (MoS2 ) = 1.2 eV (∼ 1033 nm)[310] and Eg (WS2 ) = 1.3 eV (∼ 954 nm)[310,311] ) that increases and changes to a direct band gap for single-layers (Eg (MoS2 ) = 1.8 eV (∼ 690 nm)[308,312] and Eg (WS2 ) = 1.9-2.1 eV (590-653 nm)[311,313–315] ) due to quantum confinement effects.[311,312,316] Bulk MoS2 (WS2 ) is an indirect band gap semiconductor, therefore it emits only a negligible amount of light. However, after the reduction to a single-layer sheet and thus the transition from an indirect semiconductor to a direct semiconductor the PL QY of MoS2 (WS2 ) is enhanced by a factor of around 104[308,312] (103[317] ). The charge carrier mobility of single-layer MoS2 reaches 200-700 cm2 V−1 s−1[318–320] with on/off ratios of 108 ,[319] while the highest reported value of WS2 is 50 cm2 V−1 s−1 .[321] The high charge carrier mobility and the direct band gap of TMDs allow their integration in photodetectors, light-emitting diodes, solar cells, and field-effect transistors. Single- and few-layer fabrication. There are different top-down or bottom-up methods to fabricate single- and few-layer TMD nanosheets, leading to different qualities and quantities of the resulting material. The first published method to achieve TMD layers was by chemical exfoliation using lithium intercalation, where lithium penetrates between MoS2 layers. A subsequent reaction of lithium with water produces hydrogen gas, which enables a high-yield exfoliation by ultrasonication.[301,322–324] However, due to the harsh reaction conditions, the exfoliated MoS2 layers are structurally and therefore also electronically damaged.[323] Soon after the successful chemical exfoliation of MoS2 , WS2 layers were also obtained by lithium intercalation.[324–328] Another top-down fabrication approach, which is similar to producing graphene, is the mechanical exfoliation of TMDs. The mechanical cleavage by simple rubbing or Scotch tape peeling produces high-quality single-layers, but this method is not scalable and the flake size cannot be systematically controlled.[268,308,309,317,318,329] In 2011 Coleman et al. reported on the liquid exfoliation of TMD crystals,[330,331] where the crystals were ultrasonicated in an appropriate organic or aqueous solvent, leading to large quantities of few-layer nanosheets, however, hardly achieving single-layers. In order to stabilize the exfoliated flakes against re-aggregation, polymers or surfactants (e.g., sodium cholate) have to be added to the exfoliating solvent.[331–333] Some organic solvents like NMP

31

2 Background and DMF do not necessarily need a stabilizer, because their surface energies are similar to those of the TMDs.[330,334] Another method to produce MoS2 as well as WS2 is by bottom-up fabrication with CVD. A great research interest in this technique recently lead to large-area, high-quality singe-layers by different CVD approaches.[306,313,335–346] Raman spectroscopy. Raman spectroscopy is a key technique to determine the number of layers (≤ 4 layers) in TMDs. MoS2 as well as WS2 exhibit two characteristic Raman bands, namely the E2g and A1g band. For bulk crystals these two Raman bands appear at 383 cm−1 and 408 cm−1 (MoS2 )[347] and 351 cm−1 and 420 cm−1 (WS2 ),[328] respectively (Figure 2.12). With a decreasing number of layers the frequency of the A1g band decreases, while that of the E2g band increases.[306,347] Thus, the Raman bands approach each other, which reduces the frequency difference between the E2g and A1g band. This effect becomes clearly visible from four layers downward. For more than four layers the bands in the Raman spectrum converge to the Raman bands of the respective bulk TMD.[304,347] (b)

1.5

Intensity (a.u.)

Intensity (a.u.)

(a)

A1g 1.0

E2g

0.5 0.0 360

400

Raman shift (cm-1)

440

1.5

A1g 1.0

E2g

0.5 0.0 300

400

500

Raman shift (cm-1)

Figure 2.12: Raman spectra for bulk (a) MoS2 and (b) WS2 . Both materials exhibit two characteristic Raman bands, namely the E2g and the A1g band. These bands appear at 383 cm−1 and 408 cm−1 for bulk MoS2 and 351 cm−1 and 420 cm−1 for bulk WS2 .

2.4 Semiconductor Quantum Dot Hybrids Semiconductor QDs are ideal light absorbing materials, due to their strong and spectrally tunable absorption,[37] while transport of photoinduced charges within QD networks is challenging. Many grain boundaries between the particles result in an overall low charge carrier mobility. Moreover, the concentration of free carriers is reduced due to charge carrier trapping and electron-hole recombination.[15,348] However, for most optoelectronic devices it is necessary to have an efficient and fast charge carrier separation and a subsequent charge transport to

32

2.4 Semiconductor Quantum Dot Hybrids extract the charge carriers by the electrodes. One- or two-dimensional nanomaterials like SWNTs, graphene, or TMDs exhibit charge carrier mobilities much higher than QDs,[118,263,280,320,321] but their light absorption is very weak and not freely tunable.[273,321,349] Therefore, hybrid materials of one- or two-dimensional nanomaterials sensitized with QDs are very promising for application in optoelectronic devices like solar cells or photodetectors. The QDs function as light absorbing, charge carrier generating material, while the one- or two-dimensional nanomaterial represents the conducting wire. Upon light absorption, a photoexcited state is generated in the QDs, which will be converted into a charge separated state if charge injection from the QDs to the hybrid material occurs on a timescale faster than electron-hole recombination (Figure 2.13).

Figure 2.13: Schematic illustration of the charge transfer in a QD hybrid material, exemplarily shown for PbSe-MoS2 . The incoming light is absorbed by the QD, forming an exciton. The direct, linker-free contact between the QD and MoS2 enables a fast charge separation, followed by a charge transport in the MoS2 layer.

To achieve a fast charge separation it is important to have a direct interface between the QDs and the charge transporting material without molecular linker, which, in most cases, would impede an efficient charge transfer.[350–353] Therefore it is necessary to establish a synthesis where the QDs grow directly and noncovalently on the hybrid material. A covalent functionalization of the one- or two dimensional nanomaterial should be avoided, since it induces defect sites into the charge transporting material, thus distorting its excellent charge transport properties.[354] The resulting hybrids will exhibit an efficient light absorption where the spectral response of the final material can be fine-tuned and enhanced by the QDs combined with a fast charge transport in the one- or two-dimensional nanomaterial. Another major advantage of QD hybrid materials is their costefficient solution-processability at low temperatures. Traditional materials used for solar cells and photodetectors (e.g., monocrystalline silicon or PbS) are rigid and have to be processed at high-temperatures, thus they cannot be used with

33

2 Background flexible substrates like polyethylene terephthalate (PET) or polyimide (PI), while solution-processable QD hybrid materials are suitable for lightweight, flexible electronics.[355,356]

2.4.1 Quantum Dots-Carbon Nanotubes Over the years of intense research, different ways to functionalize multi- and singlewalled carbon nanotubes with all kinds of nanoparticles were established among various groups worldwide. In this overview the focus is on metal chalcogenide QDs coupled to CNTs. All the different functionalization pathways end up in two principle strategies, that are ex situ and in situ coupling of QDs to CNTs. Ex situ coupling of QDs to CNTs. Ex situ coupling denotes the covalent or noncovalent coupling of pre-formed QDs to CNTs. In order to couple QDs covalently to CNTs, the CNTs are oxidized to create active binding sites for the ligands of the QDs. Early reports about QD-CNT hybrids are all based on covalent coupling. In 2002 publications of Banerjee et al.[357] and Haremeza et al.[358] reported on CdSe QDs coupled to oxidized SWNTs, followed by numerous publications about QD coupling to oxidized single- or multi-walled carbon nanotubes (e.g., ZnS/CdS,[359] CdS,[360] CdSe,[361] CdTe,[362] PbS,[363] and PbSe[364] ). Oxidation converts some sp2 hybridized carbon atoms into sp3 -hybridized carbon atoms, introducing scattering sites for electrons and thus affecting the inherent good electrical properties of the CNTs. The reduction of the charge carrier mobility in CNTs renders the application of these hybrids in devices less effective. The negative impact on the electrical properties is avoided by non-covalent coupling of the QDs to the CNTs. Hu et al.[365] showed that pyrene functionalized CdSe QDs can be attached non-covalently to SWNTs, followed by a publication about perylene functionalized CdSe QD-MWNT hybrids by Weaver et al.[366] The large electronic π-system of organic chromophores like pyrene or perylene connects the QDs to the CNTs by π-π-stacking forces. Another way to couple QDs non-covalently to CNTs is to make use of electrostatic interactions. CNTs are either π-π-bonded to a charged polyelectrolyte,[367] wrapped with a charged polyelectrolyte,[368,369] or wrapped with a surfactant[370] to enable coupling with oppositely charged QDs. In 2006, Guldi et al.[367] succeeded in linking CdTe QDs with negatively charged thiglycolic acid ligands to CNTs functionalized with pyrene+ . The π-system of pyrene+ binds via π-π-stacking to the CNTs, while electrostatic interactions between thiglycolic acid and pyrene+ attach the CdTe QDs to the CNTs. Another preparation method to

34

2.4 Semiconductor Quantum Dot Hybrids couple CdS QDs to SWNTs is given by Li et al.[371,372] SWNTs are non-covalently functionalized with oleylamine, where the amine group of oleylamine provides an active binding site for CdS QDs. All these methods couple QDs indirectly to CNTs and therefore increase the distance between the two materials. By increasing the distance, the charge transfer efficiency decreases. However, there are also a few rare ligands that can actually support the charge transfer efficiency between QDs and CNTs,[373] but then a laborious and time-consuming ligand exchange is necessary to get the QDs functionalized with the respective ligands. The biggest advantage of an ex situ coupling is that the QDs are pre-formed and therefore there is a good control over size, quality, and monodispersity of the QDs. In order to avoid a molecular linker in between QDs and CNTs, a direct growth or a direct deposition of the QDs onto the CNTs is essential. In situ coupling of QDs to CNTs. The term in situ coupling describes a process were QDs are either deposited or directly grown on the CNTs. Direct QD deposition is carried out by methods like laser ablation (e.g., PbS-SWNT hybrids[374] ) or CVD (e.g., CdSe-SWNT hybrids[375] ). These methods are hard to scale up and in most cases very costly due to the need of specialized equipment. Therefore, the best way to realize well performing QD-CNT hybrids is to synthesize the QDs directly, covalently or non-covalently, onto the CNTs. The first in situ growth of QDs on SWNTs was realized by Banerjee et al., where CdTe[376,377] and CdSe[378] QDs were covalently grown on oxidized CNTs. Robel et al.[379] synthesized CdSSWNT hybrids by treating SWNTs with tetraoctylammonium bromide in order to facilitate the CdS growth. However, an in situ non-covalent growth of QDs on CNTs is preferable since it does not affect the intrinsic electrical properties of the CNTs and therefore maintains the good charge carrier transport properties of the CNTs. Two good examples for an in situ non-covalent growth of QDs on CNTs were given by Ju´arez et al.[380,381] for CdSe-CNT hybrids and by Yu et al.[382] for PbSe-MWNT hybrids.

2.4.2 Quantum Dots-Graphene Graphene with its large π-system can form hybrids with all kind of nanoparticles, including metal, metal oxide, and metal chalcogenide nanoparticles. Similar to the section about QD-CNT hybrids (see Section 2.4.1), this section focuses on metal chalcogenide QD-graphene hybrids. As described in Section 2.3.1, graphene can be obtained via various methods, each method influencing its final properties. QD-graphene hybrids are reported for graphene grown by CVD and for graphene 35

2 Background exfoliated from graphite or graphite oxide, which can subsequently be reduced to graphene oxide. Reduced graphene oxide (rGO) has the advantage to show reasonable dispersibility in polar solvents, making it a good choice for wet-chemical processes. However, due to the oxidation and subsequent reduction, the carbon lattice will be damaged, limiting the hybrid performance.[288,289] Mechanically exfoliated graphene and CVD grown graphene both show excellent quality but are not processable in solution and therefore they are limited in use.[284–286] It is hardly possible to obtain single layers of graphene by using a liquid exfoliation process, however the produced few-layer graphene sheets are ideal to use in wet-chemical processes.[290] The mechanisms to produce QD-graphene hybrids are similar to the mechanisms to form QD-CNT hybrids. It is possible to couple pre-formed QDs to graphene (ex situ) and to synthesize the QDs directly onto graphene (in situ). Ex situ coupling of QDs to graphene. The most prominent way to couple pre-formed QDs to graphene is to make use of π-π-stacking forces. Therefore, the QDs are functionalized with pyridine or pyrene, which enhances the adhesion of the QDs to graphene and provides a good electronic coupling. In this way Geng et al. coupled CdSe QDs to rGO,[383] Guo et al. attached CdSe QDs to CVD grown graphene,[384] Katsukis et al. formed CdTe QD graphene hybrids using liquid exfoliated graphene,[385] while Sun et al. formed hybrids of PbS QDs and CVD grown graphene.[386] Moreover, CdTe QDs were also coupled to graphene via electrostatic interactions using pyrene+ .[385] An alternative way to connect QDs to graphene was studied by Jing et al., where methylene blue was chosen as linking molecule. While in most cases a linker molecule hampers the charge transfer from the QDs to graphene, Jing et al. claim that methylene blue can modulate the charge transfer process.[387] A simple but efficient way to form PbS-graphene hybrids was introduced by Konstantatos et al. They drop-cast pre-formed PbS QDs stabilized with oleic acid on mechanically exfoliated graphene on a Si/SiO2 substrate and subsequently performed a ligand exchange from oleic acid to the much shorter ligand ethandithiol.[352] In situ coupling of QDs to graphene. There are various reports of in situ synthesized CdS,[388] CdSe,[389–391] PbS,[392] PbSe,[355] and ZnS[393] QDs on graphene, aiming to achieve a linker-free and direct contact between QDs and graphene. CdSerGO hybrids, PbSe-electrochemically exfoliated graphene hybrids, and ZnS-rGO hybrids were synthesized via wet-chemical reactions.[355,389,393] Similar to QDCNT hybrids, there are also QD-graphene hybrids, which are formed by physical

36

2.4 Semiconductor Quantum Dot Hybrids methods. Yu et al. deposited CdSe QDs via CVD on rGO,[390] while Kim et al. deposited CdSe QDs electrochemically onto CVD grown graphene.[391] Moreover, PbS QDs were also formed on CVD grown graphene by electron beam thermal evaporation.[392]

2.4.3 Quantum Dots-Transition Metal Dichalcogenides During the time of this work, no other reports about coupling semiconducting QDs to TMD nanosheets had been published yet. However, very recently some interesting reports about QD-TMD hybrids were published. Kufer et al.[394] fabricated PbS-MoS2 hybrid photodetectors by drop-casting pre-formed PbS QDs stabilized with oleic acid on mechanically exfoliated MoS2 nanosheets on a Si/SiO2 substrate and subsequently performed a ligand exchange from oleic acid to the much shorter ligand ethandithiol. They claim that there is an exciton separation at the interface between the PbS QDs and the MoS2 layer, where the holes remain within the QD layer, while the electrons circulate through MoS2 . Prins et al.[395] took a closer look at the CdS/CdZnS-MoS2 material combination, regarding a possible energy transfer between these two materials. They found that there is a non-radiative exciton energy transfer from the QDs to monolayer and few-layer MoS2 , which quenches the QD PL QY depending on the number of MoS2 layers, with the largest quenching for monolayer MoS2 . However, in both of these reports pre-synthesized QDs were deposited ex situ on the MoS2 sheets, this is why a direct interface was not achieved. PbSe QD-MoS2 hybrid structure. An epitaxial growth of QDs on MoS2 layers would fulfill the requirement of a direct and defined interface between the two materials, avoiding a molecular linker, which usually works as charge barrier. In order to obtain a solution-processable product it is highly desirable to establish a wet-chemical synthesis, where nanoparticles grow epitaxially on the supporting nanomaterials. Hung et al. reported on a wet-chemical synthesis, where noble metal nanoparticles (i.e., Pd, Pt, Au, and Ag) were grown epitaxially on MoS2 nanosheets.[396] But so far, there are no reports on a direct growth of any kind of semiconductor QDs on TMDs. The combination of PbSe and MoS2 is considered to be promising concerning an epitaxial growth of the QDs, as it is well known that rock-salt monochalcogenides MX (M = Pb, Sn, Sb, Bi, or rare earth metal; X = S, Se, or Te) and layered dichalcogenides TX2 (T = early transition metal) form atomically sharp

37

2 Background interfaces. The material combination of PbSe and MoS2 belongs to the class of misfit-layered compounds [(MX)1−δ ]m [TX2 ]n , where two materials are alternatingly stacked.[397–400] Even if the crystal lattices of the two materials do not match in symmetry, an epitaxial relationship can be expected. These misfit-layered compounds can also be described as intercalation compounds, where the intercalation of the host crystal (TX2 ) with the intercalate MX results in new properties. The electronic features of the host structure are changed due to a charge transfer from the MX part to the TX2 layer.[401,402] Moreover, the advantage of using MoS2 over graphene is that MoS2 has sulfur atoms with lone-pairs of electrons at the surface of each layer, while graphene only offers a large electronic π-system. The lone-pairs of electrons of the sulfur atoms in MoS2 might be helpful to attract positively charged lead ions to the MoS2 surface in order to initiate Pb2+ nucleation on MoS2 and start a subsequent in situ epitaxial growth of PbSe QDs on MoS2 .

2.5 Solution-Processed Quantum Dot Optoelectronics Optoelectronic devices are a combination of electrical and optical devices. They either convert an optical signal into an electrical signal or the other way round. Optoelectronic devices include photodetectors, photovoltaics, LEDs, and lightemitting field-effect transistors (LEFETs). QDs possess a great potential to be used in a wide variety of optoelectronic devices due to their superior optical properties, like their size-dependent band gap and extremely high PL QYs. QDs exhibit low manufacturing costs and their synthesis is scalable yielding large quantities.[117] Moreover, QDs are solution-processable and therefore qualify for large-area processing both on rigid and flexible substrates. There is huge research interest in fabricating QD photovoltaics. Different research groups around the world regularly excel each other with efficiency records, with the highest power conversion efficiency of 8.55 % achieved in 2014 by Chung et al.[3,127,129,130,132,139,141,156,204,205,403–413] QD photodetectors can nowadays be fabricated with normalized detectivities around 1012 -1013 Jones, which qualifies these devices already for commercial applications. The first electroluminescence (EL) devices using QDs were reported in the late 1990s and early 2000s.[414–421] QDs are especially attractive for EL devices due to their high color purity resulting from their narrow emission bands.[422] A continuous optimization process of QD quality, processing, and device architecture[1,16,423–436] increased the external quantum ef-

38

2.5 Solution-Processed Quantum Dot Optoelectronics ficiency (EQE) of LEDs based on PbS QDs from early values of about 0.5 % in 2002[418] to EQEs of around 2 % ten years later,[437] and finally to record EQEs of 4.3 % for PbS-CdS core-shell QDs in 2015.[438] Moreover, device engineering of QD field-effect transistors (FETs) resulted in transistors with charge carrier mobilities reaching 30 cm2 V−1 s−1 .[155] High emission efficiencies and good charge transport are basic requirements to achieve light-emitting QD transistors. In the following sections the basics of photodetectors and field-effect transistors, including a literature overview on QD-based devices will be presented.

2.5.1 Photodetectors In this section the basic working principles of photodetectors, in detail photodiodes and photoconductors, will be briefly introduced, followed by an explanation of the most important figures of merit for photodetectors. Then, a literature overview on QD and QD hybrid photodetectors will be given, mainly focusing on cadmium chalcogenide and lead chalcogenide QDs. Basics of Photodetectors There are two main types of photodetectors: photodiodes and photoconductors. Photodiodes rely on the extraction of two charge carriers, while photoconductors are unipolar devices (Figure 2.14).

Figure 2.14: Schematic illustration of the band structure and charge separation process of (a) a photodiode and (b) a photoconductor. (a) Photoexcited electrons and holes are separated by the built-in potential, which occurs at the junction between a p-type (left) and a n-type (right) semiconductor through a band bending by an adjustment of the Fermi level EF . (b) In a photoconductor one type of photogenerated charge carrier is trapped (here: electrons), while the opposite charge carrier can circulate (here: holes). Depending on the lifetime of the trapped charge carrier, the opposite charge carrier can circulate many times, before it recombines with the trapped one. Figure adapted from Konstantatos et al.[92]

39

2 Background Photodiodes. Photodiodes are based on the contact of two materials with different work functions, thus producing a built-in potential. There are three combination methods of the two materials, where at least one material has to be a semiconductor. A heterojunction is formed between two different semiconductors, while one semiconductor with opposite doping levels leads to a homojunction. Moreover, the semiconductor can also be combined with a rectifying metal, resulting in a Schottky junction. The photogenerated electrons and holes are separated by the intrinsic electric field and subsequently move to the opposite electrical contacts (Figure 2.14(a)). Photodiodes show a very low sensitivity because they exhibit no photoconductive gain, unless one makes use of effects like avanlanche or carrier multiplication. A great benefit of photodiodes is their extremely fast response time. The response time is given by the transit time of the charge carriers and is therefore shorter than their lifetime, which is defined by the electron-hole recombination time.[92,348,439]

Photoconductors. A photoconductor is made of one semiconductor and two ohmic metal contacts. Upon illumination an electron-hole pair is generated, leading to a reduced resistance in the semiconductor. Photodetection is based on the trapping of one type of photogenerated charge carrier (minority charge carrier), while the opposite charge carrier can circulate (majority charge carrier). The majority charge carrier can make many cycles through the external circuit, before it recombines with the trapped charge carrier. This means that as soon as the charge carrier is withdrawn by one electrode, it is reinjected by the other electrode (Figure 2.14(b)). The lifetime of the trapped carrier has to exceed the transit time of the circulating carrier in order to achieve photoconductive gain. Photoconductors usually exhibit a high photoconductive gain, because one exciton can generate more than one charge carrier. This property renders photoconductors very sensitive devices, able to detect extremely small signals. In order to attain gain, it is important to have long living trap states to efficiently delay electron-hole recombination. However, long living trap states also lead to slow response times. Therefore, photoconductors always exhibit a trade-off between sensitivity and response time. Moreover, unlike photodiodes, photoconductors need an external electrical field to dissociate excitons.[92,348,439]

Figures of merit of photodetectors. There are different parameters that quantify the performance of photodetectors.[92,348,439] The responsivity R indicates the

40

2.5 Solution-Processed Quantum Dot Optoelectronics measureable photocurrent (IP h ) per incident optical power (Pin ) according to:[440] R=

IP h . Pin

(2.7)

The spectral response is described by the spectral dependence of the responsivity R. In QD photodetectors the spectral response curve exhibits the same shape as the QD absorption spectrum. In order to get a detectable photocurrent, the signal generated upon illumination has to be larger than the internal noise current (in ) of the detector. The noise-equivalent power (N EP ) specifies the minimum detectable optical signal by a photodetector, and is calculated according to: N EP =

in . R

(2.8)

However, N EP cannot be compared for different devices. Therefore, the most important characterization parameter is the normalized detectivity D∗ [Jones = √ cm Hz/W]. It is independent of the detector area, geometry, and bandwidth and thus allows the comparison of different photodetector devices: √ AB , D∗ = N EP

(2.9)

with the detector area A and the electrical bandwidth B.[92,439,441] Another important parameter for photoconductors is the photoconductive gain G, which is defined as the ratio between carrier lifetime (τc ) and transit time (Tt ): G=

τc . Tt

(2.10)

The photoconductive gain indicates how many cycles the photogenerated majority charge carriers can circulate before they recombine with minority charge carriers.[348] Moreover, response time, quantum efficiency as well as several noise sources should be considered when evaluating photodetector performance. The quantum efficiency indicates the number of charge carriers generated per incident photon.[440] The noise in photodetectors should be kept as low as possible. Among many noise sources, the two most important ones are Johnson noise and the noise generated by the dark current. The Johnson noise arises from the thermal motion of charge carriers, while the dark current describes the detectable leakage current when the detector is biased but not exposed to illumination.[348,440] Photodetectors with a good sensitivity can detect signals under low illumination and also show a high

41

2 Background signal-to-noise ratio under bright illumination. In order to improve a photodetector performance it is not only important to attain a large electrical output to an optical input, but also to reduce the noise.[92,348,439–441] Semiconductor Quantum Dot Photodetectors The spectral response of a QD-based photodetector can be tuned by changing the particle size, due to the size-dependent band gap of QDs. Through adjusting the spectral response to the desired application, the use of optical filters is unnecessary. However, in order to be commercially competitive, QD photodetectors need to have a detectivity between 1012 -1013 Jones and a fast response time in the tens of milliseconds.[442] The two most common semiconductor materials used for QD photodetectors are PbX and CdX (X = S, Se) QDs. Photoconductivity in QDs. Early studies of charge separation and photoconductivity in QD-polymer mixtures were published by Greenham et al. in the late 1990s.[443,444] However, the first detailed studies of photoconductivity in pure QD films were performed in the early 2000s on CdSe QD films. In this study, the photoconductivity was investigated as a function of the applied electrical field and the surface passivation of the QDs.[445–447] The measurable photoconductivity was very low, since the QDs in these early experiments were all capped with long ligands. A few years later CdSe QDs were treated with alkylamines or strong bases in order to improve interdot coupling, whereupon first charge transport studies were possible.[122,448,449] Early PbX and CdX QD photodetectors. In 2005 the first photodetector based on a pure QD film was reported by Bulovi´c et al. A CdSe QD photodiode was fabricated, which was able to detect visible light (350-575 nm) after a ligand exchange of the QD surfactant with n-butylamine.[138] In the same year the Sargent group published a QD photodetector with a responsivity of around 3 mA/W, made of a PbS QD-polymer nanocomposite. It was claimed that the photogenerated holes are transferred to the polymer, while the electrons are probably trapped or hop through the QD network.[403] In 2006 Konstantatos et al. succeeded in fabricating the first highly sensitive QD photoconductor, which was based on PbS QDs.[4] Upon illumination, the photogenerated electrons were trapped, while the holes could circulate.[92,439] The PbS QD photodetector was NIR sensitive and exhibited high gains in the 10-100 range, a bandwidth of 20 Hz, and a record detectivity of around 1.8 · 1013 Jones (at 1300 nm).[4] The presented device was

42

2.5 Solution-Processed Quantum Dot Optoelectronics able to compete with commercial InGaAs photodiodes, and therefore represented a breakthrough in QD photodetector fabrication.[26]

PbX and CdX QD photoconductors. Soon after the first high sensitive QD photodetector, the Sargent group came up with several variations of this detector.[441,442,450,451] For example, by synthesizing small PbS QDs the sensitivity of the detector was tuned into the visible wavelength range, still keeping a large photoconductive gain (> 100) and a high detectivity (D∗ > 5 · 1012 Jones at 400 nm).[441] Moreover, by taking advantage of multiple exciton generation, ultraviolet sensitive PbS QD photodetectors with responsivities of 18 A/W were achieved.[451] A different architecture was chosen by Lee et al. in 2011, where an all-inorganic QD photodetector was fabricated with CdSe/CdS core-shell QDs and In2 Se4 2 – metal chalcogenide ligands. The photodetector reached a remarkably high detectivity of more than 1013 Jones.[145] An air-stable PbS QD photodetector was reported in 2014 by Hu et al. The inorganically capped PbS QDs were passivated by a 30 nm Al2 O3 layer to avoid QD oxidation. The detectivity of these detectors reached 3.8 · 108 Jones (at 1550 nm) and was stable at operation in air for more than 60 days.

PbX and CdX QD photodiodes. The topic of QD photodiodes is not as well investigated as that of QD photoconductors, but nevertheless there are a few noteworthy reports. In 2009 a photodiode based on benzenedithiol ligand-exchanged PbS QDs was demonstrated. The PbS QD film was sandwiched between two different contacts (indium tin oxide and Al), forming a Schottky barrier. This geometry permitted a fast operation of up to 35 kHz, while still providing a high detectivity of 1012 Jones.[452] Rauch et al. decided to use a composite of PbS QDs, [6,6]-phenylC61-butyric acid methyl ester (PCBM), and poly(3-hexylthiophene) (P3HT) to fabricate a photodiode with detectivities of around 2.3 · 109 Jones at a high modulation frequency of 170 Hz. The PbS QDs were responsible for the NIR sensitization, while PCBM and P3HT were necessary for an efficient charge separation and transport.[453] Only recently a QD p-n-photodiode based on PbS QDs and either a ZnO or TiO2 film was reported. In this device, PbS QDs were the light absorbing material, and the solution-processed oxide film was responsible for the charge carrier transport. Schottky junctions normally show high dark currents, therefore the uniqueness of this p-n-junction is the combination of a low dark current (70-80 nA/cm2 ) with a high detectivity (D∗ > 1012 Jones).[454]

43

2 Background QD photodetectors without PbX and CdX. Besides PbX and CdX QDs, other semiconductors were also used to fabricate QD photodetectors. HgTe QD photoconductors were studied in order to expand the spectral sensitivity in the mid-infrared regime up to 3 µm.[6,455–458] On the other side ZnO QDs were used to fabricate ultraviolet photodetectors.[459,460] Towards a greener chemistry also heavy-metal free materials like Bi2 S3 [461] and In2 S3 [5] were investigated for visible wavelength photodetectors. Quantum Dot-Hybrid Photodetectors In QD hybrid photodetectors the charge generation is spatially separated from the charge transport. This means that the QDs are responsible for the optical response, while the second material exclusively transports the charges. This separation allows for an independent optimization of both processes.[462] In the beginning of QD hybrid photodetectors, polymers were often used as the charge separating and charge transporting material.[443,444,463–465] This approach was soon improved by the combination of QDs with fullerene derivates, since arrays of these small carbon molecules exhibit higher charge carrier mobilities than QD films.[466–468] However, the one-dimensional carbon allotrope fullerene should not remain the only carbon allotrope to form QD hybrids for photodetectors. Soon, many reports about photodetectors with hybrids of QDs and CNTs or graphene were published. Recently, also other high mobility layered materials (e.g., MoS2 ) were used to fabricate QD hybrid photodetectors.[394] QD-CNT photodetectors. The first report about photocurrent measurements in QD-CNT hybrids appeared in 2005, where Robel et al. reported of a CdS-SWNT hybrid, which showed a photocurrent response to visible light illumination.[379] This publication was followed by many reports about hybrids of SWNTs or MWNTs with CdSe,[356,365,366,380,469] CdS,[371,372] CdTe,[362] PbS,[374,470] and PbSe[364] showing photoresponse to visible and NIR light. Ju´arez et al. synthesized a CdSe-CNT hybrid, which exhibited a conductance decrease upon illumination. Therefore it was claimed that electrons tunnel from the QDs to the CNTs followed by an electron-hole recombination in the initially p-doped CNTs, leading to the detected conductance decrease. The holes remain in the QDs and are compensated via reactions with the environment. However, no systematic photodetector investigation was performed on this material.[380] One year later Hu et al. also measured a resistance increase in CdSe-pyrene-SWNT hybrids, arriving at the same conclusion as Ju´arez et al., that electrons are transferred from the CdSe QDs to the SWNTs.[365]

44

2.5 Solution-Processed Quantum Dot Optoelectronics Li et al. coupled CdS QDs non-covalently to SWNTs, using oleylamine as linker molecule. Contradictory to the former results, a current increase by over 40 % was detected under illumination, explained by n-doping of the SWNTs through oleylamine.[371,372] Gao et al. demonstrated one of the few well characterized PbSMWNT hybrid photodetectors. The device exhibits a responsivity of 583 mA/W and an extremely high detectivity of 3.5 · 1012 Jones.[470] QD-graphene photodetectors. Similar to QD-CNT hybrids, CdSe,[384,389,390] PbS,[352,386,392] and PbSe[355] QDs were also coupled to graphene in order to fabricate hybrid photodetectors. The initial report was on CdSe-rGO hybrids by Lin et al. in 2010. The hybrid shows a distinct current increase upon illumination with a 532 nm laser. However, a systematic photoresponse study is missing.[389] Guo et al. also measured a conductivity increase in CdSe-graphene hybrids upon visible light illumination, indicating an electron transfer from the CdSe QDs to graphene.[384] Better characterized photodetectors are reported for PbX (X = S, Se) QDs. Manga et al. synthesized a three component hybrid material of PbSe QDs, TiO2 nanoparticles, and exfoliated few-layer graphene, enabling an effective charge separation process. This promising material combination shows a normalized detectivity of around 3 · 1013 Jones for visible wavelengths and a slightly lower normalized detectivity of about 5.7 · 1012 Jones in the infrared region.[355] In 2012 Konstantatos et al. published a PbS-graphene hybrid photodetector with an ultrahigh gain of 108 electrons per photon and a normalized detectivity of 7 · 1013 Jones. The extremely high gain results from a combination of the high charge carrier mobility of graphene and the long living trap states in PbS QDs.[352]

2.5.2 Field-effect Transistors This section introduces the basics of unipolar and ambipolar FETs, including their working principle and the most important equations. Then, the concept of electrolyte-gating will be explained, followed by the introduction of LEFETs. At the end of this section a literature overview on QD-based transistors will be given. Basics of Field-effect Transistors FETs are purely voltage controlled electrical switches. Typical FETs consist of three electrodes, i.e., source, drain, and gate electrode,[348] a dielectric layer, and a semiconducting layer. The source and drain electrode of a defined channel width W are separated by a distance L (channel length) forming the transistor

45

2 Background channel (Figure 2.15).[471] The semiconductor is located in the transistor channel and is contacted by the source and the drain electrode. The thin dielectric layer electrically insulates the gate electrode from the semiconductor in the transistor channel. The dielectric layer can be an inorganic insulator (e.g., an oxide like SiO2 , Al2 O3 , HfOx ), a polymeric insulator (e.g., poly(methyl methacrylate) (PMMA)), or an electrolyte (e.g., ionic liquid).[471] Usually, the gate and drain electrodes are biased, while the source is grounded. The charge carrier flow and therefore the conductance in the transistor channel is generated by the applied voltage between source and drain (Vd ) and can be modulated by the gate voltage (Vg ).[348]

Figure 2.15: Schematic illustration of a top-gated FET structure, depicting the position of the source, drain, and gate electrode, the substrate, the semiconductor, and the dielectric. L and W represent the channel length and the channel width. The source electrode is grounded, while gate and drain electrode are biased. Figure adapted from Zaumseil et al.[471]

In principle, there are unipolar FETs made of predominantly electron (n-type) or hole (p-type) conducting semiconductors as well as ambipolar FETs, which are fabricated with semiconductors that can equally conduct holes and electrons. In a unipolar transistor the source is the charge-injecting electrode for both electrons (Vg > 0) and holes (Vg < 0), whereas in an ambipolar transistor source and drain electrode should be named electron source and hole source, since both electrodes are able to inject charge carriers into the semiconductor.[471] Most transistors operate in the enhancement-mode, therefore they are normallyoff. If the gate is not biased, the channel conductance is very low. As soon as a gate voltage is applied, mobile charge carriers will be accumulated at the semiconductor/dielectric interface and the channel will become conductive. However, there are also transistors which are normally-on with no gate bias. These transistors operate in the depletion-mode, where a gate bias has to be applied to decrease the channel conductance and turn the transistor off.[440] A transistor can be fabricated with various device geometries, resulting in

46

2.5 Solution-Processed Quantum Dot Optoelectronics different device behavior. The three most popular architectures are bottom contact/top gate, bottom contact/bottom gate, and top contact/bottom gate. A more recent device design is the bottom contact/side gate structure. All these architectures mainly differ in their relative position of the gate electrode to the source and drain electrode.[348,471] In this work a bottom contact/side gate structure was used. This structure is compatible with ionic liquids and ion gels as dielectrics. Moreover, there is no need for an additional gate evaporation step since the gate electrode can be patterned at the same time as source and drain electrode.

Figures of merit of FETs. In order to characterize and compare different FETs, it is important to define some figures of merit. These include the charge carrier mobility µ, the threshold voltage Vt , the subthreshold swing, the ON/OFF ratio, the gate leakage, and the hysteresis. The charge carrier mobility µ in FETs is given by vd = µ x Ef and denotes how fast charge carriers can travel through the semiconductor (given by the drift velocity vd ) when a certain electric field Ef is applied.[440] The threshold voltage Vt defines the necessary gate voltage to switch a transistor from the OFF to the ON state. The subthreshold swing characterizes the gate voltage, which is necessary to cause a change of the drain current of one order of magnitude.[440] The ON/OFF ratio is the ratio between the maximum drain current in the on-state and the minimum drain current in the off-state. The maximum drain current is mainly determined by the charge carrier mobility and the dielectric capacitance, assuming the contact resistance is negligible. The minimum drain current depends on the gate leakage, this is why it should be kept as small as possible. The gate leakage denotes the current from the gate to the source or drain electrode, and decreases with increasing quality of the dielectric layer. A high ON/OFF ratio is desirable because it ensures a clear switching behavior and therefore indicates a high-performance transistor.[471] The hysteresis can be extracted from the difference in forward and reverse drain current and is usually explained by charge carrier trapping. Ideally, both curves resemble each other.[155]

Metal-insulator semiconductor capacitor. If no drain bias is applied, a FET can be seen as a metal-insulator semiconductor capacitor. The capacitance Ci of the dielectric layer is then given by:[348,440] Ci =

Aεε0 W Lεε0 = , dDL dDL

(2.11)

47

2 Background where A is the area of the capacitor, defined by the transistor channel length L and width W (A = W x L), ε is the relative dielectric constant, ε0 is the permittivity of vacuum, and dDL is the thickness of the dielectric layer. When a gate voltage Vg is applied, a certain concentration of charges (either electrons for Vg > 0 or holes for Vg < 0) will be accumulated at the insulator/semiconductor interface. The amount of accumulated charges is directly proportional to the gate voltage and the capacitance of the insulator. However, some of these charges will be trapped until the threshold voltage Vt is reached and all trap states are filled, leading to an effective gate voltage of Vg − Vt . Therefore, mobile charges will only start to accumulate after reaching the threshold voltage. The amount of mobile charges Qmob per unit area is:[471] Qmob = Ci (Vg − Vt ).

(2.12)

Thus, when Vg is constant, there are two ways to increase the amount of mobile charges Qmob injected in a FET channel: Either the dielectric thickness dDL has to be reduced or a dielectric with a high dielectric constant ε has to be chosen.[348]

Unipolar FETs. Applying a voltage to the gate electrode will generate an electric field across the dielectric, and thus perpendicular to the channel. This electric field accumulates oppositely charged mobile carries at the semiconductor-dielectric interface. In a unipolar FET usually one charge carrier (majority charge carrier) is preferentially accumulated at the interface compared to the opposite charge carrier (minority charge carrier), which either leads to a n- or p-channel.[472] In n-type FETs (and vice versa for p-type FETs), where electrons are the majority charge carriers, a positive gate voltage enhances the free charge carrier density and therefore increases the current flow between source and drain, while a negative gate voltage depletes the transistor channel and thus decreases the channel conductance.[473] When the gate voltage approaches the threshold voltage Vt the transistor channel switches from insulating to conducting. Therefore by applying different gate voltages, the transistor can be switched on and off.[348] For small drain voltages Vd (Vd  (Vg − Vt ) and Vg > Vt ), the transistor operates in the linear regime, having a linear potential drop in the channel from the drain electrode (V = Vd ) to the source electrode (V = 0). The drain current Id increases linearly with Vd (Figure 2.16(a)): Id,lin =

48

Vd W Ci µlin (Vg − Vt − )Vd . L 2

(2.13)

2.5 Solution-Processed Quantum Dot Optoelectronics Vd/2

is negligible, because Vd is small, which leads to:[348,440,471] W Ci µlin (Vg − Vt )Vd , L

Id,lin =

(2.14)

where W is the channel width, L the channel length, µlin the linear charge carrier mobility, and Ci the capacitance per unit area of the dielectric layer. The gradient of the drain current Id versus the gate voltage Vg at a constant drain voltage Vd yields the linear charge carrier mobility µlin :[471] µlin =

∂Id,lin L . ∂Vg W Ci Vd

(2.15)

By increasing the drain voltage Vd up to Vd = (Vg − Vt ) and keeping Vg still above Vt , the channel pinches-off next to the drain electrode (Figure 2.16(b)). When Vd is further increased (Vd > (Vg − Vt )), the depletion region expands and the pinch-off point moves through the channel towards the source, causing a shortening of the transistor channel. Moreover, the drain current saturates, which means it is independent of Vd and cannot increase further (Figure 2.16(c)). Substituting Vd in equation 2.13 with (Vg − Vt ) yields the saturation current Id,sat :[348,440,471] Id,sat =

W Ci µsat (Vg − Vt )2 , 2L

(2.16)

where µsat is the saturation charge carrier mobility that can be calculated as follows:[471] µsat =

1 ∂Id,sat L . ∂Vg W Ci (Vg − Vt )

(2.17)

In QD transistors the saturation mobility often exceeds the linear mobility, due to an initial charge carrier trapping until most of the trap states are filled. Ambipolar FETs. An inevitable condition to fabricate ambipolar transistors is a channel material which has similar electron and hole conduction and is therefore able to accumulate both charge carriers.[472] Moreover, in order to inject both, electrons and holes, the electrode materials have to be chosen in a way that the work function of one metal aligns with the VB for hole injection whereas the work function of the second metal has to line up with the CB for electron injection. Since lead chalcogenide semiconductors have a small band gap (see Section 2.1.4), both charge carriers can easily be injected from only one electrode material without a large contact resistance for neither electrons nor holes.[471,472] Depending on biasing conditions, electrons and holes are injected into the channel simultaneously from 49

2 Background (a)

G

Vg > Vt

Semiconductor

Id

Dielectric S

Channel potential Substrate

D Vd Vt

S

Id D Vd = Vg – Vt Vd

(c)

G S

Vg > Vt

Pinch-off point Id D Vd > Vg - Vt Vd

Figure 2.16: Schematic illustration of the channel potential distribution between source and drain electrode and the corresponding output curves for (a) the linear regime, (b) the pinch-off point, and the (c) saturation regime. Figure adapted from Zaumseil et al.[471]

source and drain electrodes. As depicted in Figure 2.17, ideal ambipolar transfer curves exhibit a characteristic V-shape with three regimes. If the gate voltage Vg is kept constant and the drain voltage Vd is increased, the potential difference between drain and gate electrode changes, leading to a shift of the transfer curve and therefore a spatial shift of the ambipolar region in the transistor channel. By sweeping the gate voltage Vg and keeping the drain voltage Vd constant, the majority charge carrier is changed, which becomes clear in the V-shaped transfer curve. The curve changes from the unipolar electron accumulation regime (n-type) for positive Vg (i.e., Vg ≥ Vt,e and Vd ≤ (Vg − Vt,e )) via the ambipolar regime (i.e., Vg > Vt,e and Vd ≥ (Vg − Vt,h )) to the unipolar hole accumulation regime (p-type) for negative Vg (i.e., Vg < Vt,e and Vd ≥ (Vg − Vt,h )).[348] From the unipolar drain currents (see equation 2.13 and 2.16) the ambipolar drain current Id,ap can be derived as follows:[471,474] Id,ap =

WC {µe (Vg − Vt,e )2 + µh (Vd − Vg + Vt,h )2 }. 2L

(2.18)

Ambipolar transistors do not exhibit a real off state, because they switch from one unipolar to the other unipolar regime via the ambipolar regime. This is why ambipolar transistors are poor switches. However, due to their ability to 50

2.5 Solution-Processed Quantum Dot Optoelectronics

Drain Current (A)

transport electrons and holes, they can be used as LEFETs or in complementarylike circuits.[471] hole transport

electron transport

ambipolar transport

Gate Voltage (V) Figure 2.17: Characteristic V-shaped ambipolar transfer curve, with unipolar hole accumulation regime (“red”), ambipolar regime (“green”), and unipolar electron accumulation regime (“blue”).

Electrolyte-gating Electrolyte-gating is a promising technique to develop high-performance FETs with many different solution-processable semiconductors.[475] In electrolyte-gated transistors the classical inorganic gate dielectric is replaced with an electrolyte.[476] The electrolyte consists of mobile cations and anions and is therefore an ionic conductor but electrical insulator.[472] If the gate is not biased the ions are randomly distributed within the electrolyte (Figure 2.18(a)). However, if a voltage is applied to the gate electrode the resulting electric field causes a motion of the ions. Depending on the polarity of the gate bias, ions (either cations or anions) will accumulate at the semiconductor/electrolyte interface while the counter ions will move to the electrolyte/gate interface, forming two electric double layers (EDLs) (Figure 2.18(b)).[472,476] The applied voltage almost entirely drops across the few nanometer thick EDLs (d ∼ 2 nm), while the bulk electrolyte remains charge neutral. According to equation 2.11 a small dDL causes very high effective capacitances.[476] Ionic liquid based electrolytes have intrinsic capacitances of 1-10 µFcm−2 , which is much higher than capacitances of commonly used high-ε dielectrics like Ta2 O5 or HfO2 , because these oxides have a minimum thickness of around 10-100 nm.[471,472,477] The high capacitance of ionic liquids induces the accumulation of very high charge carrier concentrations (∼ 1014 cm−2 ) at very low operation voltages (< 3 V).[476] Due to these extremely high charge carrier densities, all trap states are filled. Therefore, the minority charge carrier becomes mobile in the transistor channel, resulting in transistors with ambipolar characteristics.[478] Electrolyte-gating is especially useful for nanostructured materials like QDs because

51

2 Background the electrolyte penetrates the porous network and forms a wrap-around gate for every QD.[479] However, electrolyte-gated transistors exhibit a quite high gate leakage due to ionic conductivity and a limited switching speed due to slow ionic diffusion.[480,481] Electrolyte-gating can be performed with different materials, including electrolyte solutions, polymer electrolytes, or ionic liquids.[472,482,483] Ionic liquids are molten salts, which exhibit a melting point below 100 ◦ C and are mainly liquid at room temperature.[484,485] Liquids are not suitable for application in integrated circuits, therefore nowadays mostly solid yet flexible ion gels are used as dielectrics. Iongels are formed by the integration of an ionic liquid in a block copolymer matrix. Within this matrix it is important that the ions remain mobile in order to keep the good ionic conductivity of the ionic liquids. In iongels the high ionic conductivity is combined with the good electronic insulation and flexibility of polymers, which renders iongels ideal for printable electronics.[475,486]

Figure 2.18: Schematic illustration of the ion distribution of an electrolyte-gated transistor. (a) If no voltage is applied to the gate electrode, the ions are randomly distributed. (b) By applying a gate voltage, an electric field is induced, which leads to a motion of the ions. In this example, the anions accumulate at the electrolyte/gate interface and the cations accumulate at the semiconductor/electrolyte interface, forming two EDLs. Figure adapted from Jakubka.[223]

Light-emitting Field-effect Transistors For commercial semiconductor technology it is important to integrate electronic and optical functions in one device, in order to save space in an integrated setup. LEFETs provide such a combination, by integrating the switching functionality of a FET and the emission properties of a LED in one device.[471,472] To achieve light-emission from the transistor channel it is necessary to fabricate FETs with ambipolar materials, to be able to simultaneously have electrons and holes in the transistor channel. Subsequently, the separate electron and hole accumulation layers can meet within the channel, leading to an efficient electron-hole recombination. This recombination results in light emission from a narrow line (the recombination

52

2.5 Solution-Processed Quantum Dot Optoelectronics zone) in the channel (Figure 2.19(a)) and is called electroluminescence (EL). EL spectra, generated by electrical stimulation, usually have a similar spectral shape as PL spectra, which are generated by an optical excitation.[471,487,488] As introduced before the ambipolar regime and therefore the light emission zone can be shifted arbitrarily through the entire transistor channel by varying the applied gate and drain voltage (Figure 2.19(b)). If the emission zone is not located at the electrodes but somewhere within the channel this is a clear evidence for truly ambipolar transport and thus the coexistence of an electron and hole accumulation layer. In LEFETs there is no need for electron or hole blocking layers, because at the meeting point of the electron and hole accumulation layer all injected electrons and holes must recombine, which implies that they are always perfectly balanced. Moreover, there are no quenching effects from the metal electrodes, therefore the emission efficiency should be the highest in the middle of the channel.[471,487,488] (b)

Iongel

Electrons

Changing Vg

Drain Holes

Constant Vd

S

Source

Charge Carrier Density

(a)

D

G

Iongel S

D

G

Iongel S

D

G

Channel Lenght Figure 2.19: (a) Schematic illustration of the electron and hole charge carrier density in a transistor channel, with light emission (recombination zone) at the meeting point of the opposite charge carriers. (b) Schematic illustration of the spatially moving recombination zone from the source to the drain electrode at a changing gate voltage. Figure (a) adapted from Jakubka.[223]

From a scientific point of view LEFETs are very interesting in order to study and understand the electrically stimulated recombination and emission dynamics at high carrier concentrations and compare the transport of two opposite charge carriers in one material. Unlike classical LEDs, which consist of planar functional layers sandwiched between anode and cathode, LEFETs exhibit a lateral device architecture with an easily accessible recombination zone.[471,487] Moreover, due to their ability to accumulate very high charge carrier densities, all trap states in the semiconductor will be filled, which provides access to intrinsic semiconductor properties.[472] Under very high electric fields, EL can also be seen in unipolar FETs. This can be explained by the tunneling of minority charge carriers from an electrode in the semiconducting layer. However, in this case, the emission will always be locally fixed at the edge of one electrode and can never be moved in the

53

2 Background middle of the channel.[487–490] The first LEFETs with a spatially moveable emission zone, based on ambipolar transport in a single semiconductor, were reported by Zaumseil et al. and Swensen et al.[488,491] In the following years, ambipolar LEFETs have been demonstrated for a wide variety of different materials like bulk organic semiconductors (conjugated polymers[487] and single crystals[492] ), two-dimensional semiconductors (e.g., TMDs like WS2 [493] or MoS2 [494] ) and onedimensional semiconductors (single CNTs[495] and CNT networks[496] ). However, LEFETs based on zero-dimensional semiconductors have not been demonstrated so far, probably due to an insufficient ambipolar charge transport or a strong Auger quenching of emission in QD thin-films. Electrolyte-gating may help to accumulate very high charge carrier densities in QD thin-films and thus to achieve light emission from the channel. Semiconductor Quantum Dot Field-effect Transistors The use of semiconducting QDs in FETs is not only interesting in order to find a solution-processable, easy to synthesize and reasonably stable alternative to organic semiconductors but also to study fundamental charge transport behavior in QD films. Basic requirements for the fabrication of high-mobility FETs are a very homogeneous QD size and shape distribution as well as few surface and crystal defects.[472] QD solids have to be optimized regarding interparticle coupling and concentration of mobile carriers (see Section 2.1.3), to yield well performing FETs. Moreover, dielectric (e.g., SiO2 , Al2 O3 , PMMA, ionic liquids etc.)[119,135,152,497] and electrode (e.g., Au forms an almost Ohmic contact to lead chalcogenides[348,463,498] ) materials are two more parameters, which can be changed in order to tune FET performance and reach commercially competitive devices. Early use of QDs in FETs. The processing of semiconductor QDs in FETs started in 1999, when Ridley et al. fabricated a transistor with an annealed CdSe QD film and detected charge carrier mobility. However, due to sintering of the CdSe QDs to a polycrystalline film during annealing, the FET cannot really be seen as a QD FET, because during operation QDs were no longer present.[158] All early QD FET investigations were based on cadmium chalcogenide QDs.[136,159,447,449] Although some optimization efforts were done, film conductivity and gate modulation remained problematic. Beginning of QD FETs. In 2005 a milestone in QD FET research was presented by Talapin and Murray, when the first FET based on a non-annealed QD film

54

2.5 Solution-Processed Quantum Dot Optoelectronics was demonstrated. In order to achieve sufficient charge carrier transport and gate modulation, the PbSe QDs were treated with hydrazine, which decreased the interdot spacing from ∼ 1.2 nm to ∼ 0.3 nm. The hydrazine treatment increased the film conductance around ten orders of magnitude, resulting in an electron mobility of 0.9 cm2 V−1 s−1 and a hole mobility of 0.2 cm2 V−1 s−1 .[119] This breakthrough result started a wide interest in QD FET research, with the two most important goals of improving charge carrier mobility and switching speeds.[116–118,123,135,145,146,152,153,155,177,178,188,478,497–508] In 2009 Kang et al. processed the first ionic liquid gated QD FET with hydrazine treated PbSe QDs. The so fabricated device was slightly ambipolar but predominantly n-type and reached electron mobilities of 0.4 cm2 V−1 s−1 and hole mobilities of 0.02 cm2 V−1 s−1 .[497] A few years later an ionic liquid gated FET based on ligand-exchanged PbS QDs was realized, exhibiting charge carrier mobilities of one order of magnitude higher (µe = 1.91 cm2 V−1 s−1 ; µh = 0.15 cm2 V−1 s−1 ) than QD-based transistors before.[478] Another approach to improve QD FET performance was demonstrated by Koh et al. using PbS nanocubes instead of spheres. These nanostructures still show quantum confinement effects but compared to spheres they form closer packed films with less void volume and therefore they exhibit a stronger interparticle coupling energy.[153]

High-mobility QD FETs. After the first well-working QD FETs, a second landmark was set in 2011 by Lee et al. with the use of inorganic ligands forming all-inorganic QD nanocomposite FETs.[145] CdSe FETs with metal chalcogenide complex ligands (In2 Se4 2 – ) reached extremely high linear mobilities between 1016 cm2 V−1 s−1 and a remarkable stability over many switching cycles.[145] This fully inorganic approach was optimized, for example through trap filling by indium[118] or/and using ultrasmall inorganic ligands like S2 – [155] and thiocyanate,[118] nowadays achieving record mobilities of 10-30 cm2 V−1 s−1 for CdSe QDs,[118,145,152,155] 5-10 cm2 V−1 s−1 for PbSe QDs,[117,506] and 10-15 cm2 V−1 s−1 for InAs.[146,177] It was found that the surface atoms have a great influence on the electronic properties of a QD FET. The Fermi level can be changed via the surface stoichiometry, shifting a transistor from p- to n-type and vice versa.[117,504,506] In PbX (X = S, Se) QD FETs, a lead rich QD surface leads to n-type behavior, while a chalcogenide rich surface results in p-type behavior.[117,506] The Law group published work on air-stable PbS and PbSe QD FETs with atomic layer deposited AlOx , which serves as passivating and protecting layer. The AlOx infills the QD solid and thus suppresses QD oxidation and ripening, leading to air-stable ambipolar QD

55

2 Background FETs.[178,503,505] All-inorganic QD solids with very small interdot spacing were achieved by using the ultrasmall ligand S2 – . FETs with these all-inorganic QD solids reached high, air-stable electron mobilities of 7 cm2 V−1 s−1 . In addition, the major carrier type can be controlled with atomic layer deposition (ALD). The deposited material passivates electron traps and acceptors at the QD surface and therefore the number of ALD cycles can tune the transistor behavior from p-type via ambipolar to n-type.[505] Through a systematic tuning and optimization of a wide variety of parameters, deep insights in charge transport in QD solids are gained (see also Section 2.1.3). QD FETs are on a good way to fulfill all necessary requirements for commercial application, which are high carrier mobilities,[145] low operational voltages,[478] small hysteresis,[155] and fast switching speeds (> 10 kHz[155] ). It was already possible to fabricate complex QD integrated circuits, to realize inverters, amplifiers, and ring oscillators.[152] Moreover, QD FETs were produced on rigid as well as flexible substrates, which additionally broadens their application versatility.[152]

56

3 Experimental Part

This chapter introduces all experimental procedures that were used to obtain the results of this thesis. It describes how PbS quantum dots, PbSe quantum dots, and PbSe hybrids were synthesized. Then it discusses the fabrication process of quantum dot-based photodetectors and light-emitting field-effect transistors. The next part shortly describes the theoretical background of all methods used to characterize the synthesized materials and explains which parameters, equipment, and calculations were applied. The last part comments on the parameters, equipment, and calculations that were applied to extract figures of merit from the photodetector and light-emitting field-effect transistor measurements.

3 Experimental Part

3.1 Chemicals All chemicals were obtained from commercial sources and used as received. CoMoCAT single-walled carbon nanotubes (SWNTs) (SouthWest Nano-Technologies, Inc.; diameter 0.7-1.1 nm), MoS2 powder (particle size < 2 µm), WS2 powder (particle size < 2 µm), graphite flakes (particle size +100 mesh; ≥ 75 % min), lead oxide, lead(II) acetate trihydrate (Pb(COOCH3 ) · 3H2 O), tri-n-octylphosphine (TOP), selenium powder, diphenyl ether (≥ 99 %), bis(trimethylsilyl) sulfide (TMS), 1-octadecene (technical grade 90 %), oleic acid (technical grade 90 % for PbSeSWNT synthesis; ≥ 99.0 % for PbSe-layered material hybrids and PbS synthesis), 3-mercaptopropionic acid (MPA) (≥ 99 %), poly(vinylidene fluoridecohexafluoropropylene) (P(VDF-HFP)) (MW ∼ 400 kg·mol−1 , Mn ∼ 130 kg·mol−1 ), octane (anhydrous, ≥ 99 %), hexane, toluene, n-butanol, methanol, and acetone were purchased from Sigma Aldrich, ethanol was purchased from Carl Roth GmbH, the ionic liquids 1-ethyl-3-methyl-imidazolium tris(pentafluoroethyl)trifluorophosphate ([EMIM][FAP]) (high purity grade, dried in vacuum) and 1-ethyl-3-methyl-imidazolium bis(trifluormethylsulfonyl) imide ([EMIM][TFSI]) (high purity grade, dried in vacuum) were purchased from Merck, poly(methyl methacrylate) (PMMA) (high purity, electronic grade, MW = 300 kg·mol−1 , PD = 1.05) was purchased from Polymer Source Inc., tetrakis(dimethylamino)hafnium (TDMAH), 98+ % (99.99+ %-Hf, < 0.2 % Zr) PURATREM was obtained from Strem Chemicals.

3.2 Syntheses All performed syntheses are based on the principles of a standard quantum dot (QD) hot-injection synthesis,[46] which was adapted and further developed to yield the hybrid compounds. For details on hot-injection synthesis principles, see Chapter 2, Section 2.1.1.

3.2.1 PbSe Quantum Dots PbSe QDs for comparison to PbSe-SWNT hybrids. PbSe QD synthesis was performed under argon atmosphere using a standard air-free Schlenk line technique and the hot-injection method (Figure 3.1). For a typical synthesis (Pb:Se = 1:1) a stock solution of Pb(oleate)2 was prepared by dissolving PbO (225 mg) and oleic acid (1 mL) in 1-octadecene (7.5 mL). The reaction mixture was dried under vacuum

58

3.2 Syntheses at 100 ◦ C for 1 h. It was then heated to 120 ◦ C under a moderate argon flow, followed by an injection of diphenyl ether (5 mL) into the Pb(oleate)2 stock solution. The solution was again dried at 100 ◦ C for 10 min. It was then heated to 170 ◦ C under a moderate argon flow and a TOPSe (0.8 mL; 100 mg Se/1 mL TOP) solution was injected rapidly. The reaction mixture was kept at 150 ◦ C for the required growth time and was then quenched with a cold water bath. The PbSe QDs were purified two times with hexane/acetone and hexane/ethanol/methanol by centrifugation at 6000 rpm for 5 min to remove all contaminations and unreacted precursors. The PbSe QDs were then re-dispersed in hexane for further measurements. In order to test the influence of growth time and Pb to Se ratio, syntheses with different growth times (60 s, 5 min, 24 h) and different Pb to Se ratios (1:1, 1:2, 2:1) were performed. P mbar

manometer

Ar 2 5 0. 0 170°C J-KEM Scientific

vacuum vacuum

MOTOR

DEW AR dewar

sulfur or selenium precursor thermocouple

lead precursor

pump on

Figure 3.1: Schematic illustration of a common synthesis setup to perform a hot-injection QD synthesis under inert atmosphere.

PbSe QDs for post-synthesis mixing with MoS2 nanoflakes. PbSe QD synthesis was performed under argon atmosphere using a standard air-free Schlenk line technique and the hot-injection method. Pb(oleate)2 was prepared by dissolving PbO (225 mg) and oleic acid (1 mL) in 1-octadecene (7.5 mL). The reaction mixture was dried under vacuum at 100 ◦ C for 1 h. It was then heated to 170 ◦ C under a moderate argon flow and a TOPSe (1.8 mL; 438.3 mg Se/5 mL TOP) solution was injected rapidly. The reaction mixture was kept at 150 ◦ C for 5 min and was then quenched with a cold water bath. The PbSe QDs were purified two times with hexane/acetone and hexane/ethanol/methanol by centrifugation at 6000 rpm for 5 min.

59

3 Experimental Part

3.2.2 PbS Quantum Dots Synthesis based on lead(II) acetate trihydrate and TMS. PbS QD synthesis was performed under argon atmosphere using a standard air-free Schlenk line technique and the hot-injection method. For a typical synthesis a stock solution of Pb(oleate)2 was formed by dissolving Pb(COOCH3 ) · 3H2 O (758.6 mg) and oleic acid (6-10 mL, depending on QD size) in 1-octadecene (12 mL). The reaction mixture was dried in vacuum at 100 ◦ C for 2 h. It was then heated to 145 ◦ C under moderate argon flow. The sulfur precursor solution (210 µL of TMS in 6-10 mL 1-octadecene, prepared in a glovebox) was injected rapidly into the Pb(oleate)2 stock solution and the temperature was decreased to 120 ◦ C. The reaction mixture was kept at 120 ◦ C for 10 min and then quenched with a cold water bath. The PbS QDs were purified two to three times with a mixture of hexane/ethanol/acetone by centrifugation at 6000 rpm for 3 min to remove all contaminants. Subsequently, the PbS QDs were transferred into a dry nitrogen glovebox, redispersed in ∼ 4 mL octane and centrifuged to remove agglomerated and undissolved QDs. The supernatant was kept and used for characterization and device fabrication. It was possible to control the QD size by varying the oleic acid concentration, because oleic acid essentially influences the nucleation step. Oleic acid acts as capping ligand and therefore the more oleic acid is present, the smaller is the number of nuclei and the initial nuclei size, leading to a larger final PbS QD size.[47,509]

Synthesis based on lead oxide and elemental sulfur. PbS QD synthesis was performed under argon atmosphere using a standard air-free Schlenk line technique and the hot-injection method. For a typical synthesis, a sulfur stock solution and a lead stock solution were prepared. The sulfur stock solution was made by dissolving sulfur (40 mg) in oleylamine (7.5 mL). This solution was dried in vacuum at 100 ◦ C for 30 min and then heated to 120 ◦ C under a moderate argon flow until all sulfur was dissolved. Before further processing, it was left to cool down to room temperature under a moderate argon flow. The lead stock solution was formed by dissolving PbCl2 (2.50 g) in oleylamine (7.5 mL). This mixture was dried in vacuum at room temperature for 20 min, and subsequently heated to 120 ◦ C under a moderate argon flow. When the temperature reached 110 ◦ C, argon was turned off and the solution was degassed another 10 min. After degassing, a moderate argon flow was re-adjusted and the sulfur precursor solution (2.25 mL) was rapidly injected as soon as the lead stock solution reached 120 ◦ C (Pb:S ratio 1:24). The

60

3.2 Syntheses reaction mixture was kept at 120 ◦ C for 2.5-7 min (depending on the final QD size) and was then quenched by injecting cold hexane (20 mL). Unreacted educts were allowed to settle for one day before the reaction mixture was centrifuged one to two times at 8000 rpm for 10 min. Afterwards, n-butanol and methanol were added to the supernatant and the whole dispersion was centrifuged at 6000 rpm for 3 min. The precipitated PbS QDs were dried with a nitrogen gun and purified with hexane (6000 rpm, 3 min) to remove aggregates. Then, oleic acid was added to perform a solution-based ligand exchange from oleylamine to oleic acid. Subsequently, QDs were purified one more time with n-butanol and methanol (6000 rpm, 3 min). In order to be sure to remove all educts and side-products, the hexane, oleic acid, n-butanol, and methanol purification steps were repeated once more. Finally, PbS QDs were dried with a nitrogen gun, re-dispersed in octane and centrifuged to remove agglomerated and undissolved QDs. The supernatant was kept and used for characterization and device fabrication.

3.2.3 PbSe-Single-walled Carbon Nanotube Hybrids PbSe-SWNT hybrid synthesis was performed under argon atmosphere using a standard air-free Schlenk line technique and the hot-injection method. For a typical synthesis (Pb:Se = 1:1) a stock solution of Pb(oleate)2 was prepared by dissolving PbO (225 mg) and oleic acid (1 mL) in 1-octadecene (7.5 mL). The reaction mixture was dried under vacuum at 100 ◦ C for 1 h. It was then heated to 120 ◦ C under a moderate argon flow. CoMoCAT SWNTs (5 mg) were dispersed in diphenyl ether (5 mL) by 1 h of ultrasonication. The SWNT dispersion was degassed at 60 ◦ C for at least 30 min and was then kept under a moderate argon flow. The degassed SWNT dispersion was injected into the Pb(oleate)2 stock solution, and the mixture was again dried at 100 ◦ C for 10 min. The mixture was then heated to 170 ◦ C under a moderate argon flow and a TOPSe (0.8 mL; 100 mg Se/1 mL TOP) solution was injected rapidly. The reaction mixture was kept at 150 ◦ C for the required growth time and was then quenched with a cold water bath. The PbSe-SWNT hybrids were purified four times with hexane/acetone, hexane, hexane/ethanol/methanol, and hexane by centrifugation at 6000 rpm for 10 min to remove all contaminations and unbound PbSe QDs. The PbSe-SWNT hybrids were re-dispersed in hexane for further measurements. In order to test the influence of growth time and Pb to Se ratio, several syntheses with different growth times (60 s, 5 min, 24 h) and different Pb to Se ratios (1:1, 1:2, 2:1) were performed.

61

3 Experimental Part

3.2.4 PbSe-Layered Material Hybrids MoS2 , WS2 , and graphite exfoliation. MoS2 and WS2 powders were used as received. The graphite powder was sieved through a 0.5 mm mesh to remove large particles before exfoliation. MoS2 , WS2 , or graphite (20 mg) was dispersed in diphenyl ether (10 mL) by sonication in an ultrasonic bath (Branson 2510) for 1 h followed by vigorous sonication using a tip sonicator (Sonics Vibra Cell) running at 25 % of maximum power for 10 min. After centrifugation at 500 rpm for 45 min (Beckman Coulter Avanti J26XP centrifuge) the dispersion was allowed to settle for 2 h. The supernatant contained exfoliated MoS2 flakes, exfoliated WS2 flakes, or few-layered graphene (FLG). Subsequently, 5 mL of the respective supernatant were used for the following hybrid synthesis. Hybrid synthesis. PbSe-layered material hybrid synthesis was performed under argon atmosphere using a standard air-free Schlenk line technique and the hotinjection method. For a typical synthesis a stock solution of Pb(oleate)2 was prepared by dissolving PbO (225 mg) and oleic acid (1 ml) in 1-octadecene (7.5 mL). The reaction mixture was dried in vacuum at 100 ◦ C for 1 h, and subsequently heated to 120 ◦ C under a moderate argon flow. 5 mL of the respective nanoflake dispersion (MoS2 , WS2 , or FLG) was injected into the Pb(oleate)2 stock solution and the mixture was again dried at 100 ◦ C for about 30-60 min. It was then heated to 170 ◦ C under a moderate argon flow and a TOPSe (1.8 mL; 438.3 mg Se/5 mL TOP) solution was injected rapidly. The reaction mixture was kept at 150 ◦ C for 5 min and was then quenched with a cold water bath. The PbSe-layered material hybrids were purified four times with hexane, hexane/ethanol/methanol, hexane, and hexane by centrifugation at 10.000 g for 15 min to remove all contaminants and unbound PbSe QDs. Finally, hybrids were re-dispersed in hexane for further measurements.

3.3 Device Fabrication In the present work two kinds of optoelectronic devices were used. Photodetectors, which convert an optical signal into an electrical signal and light-emitting field-effect transistors (LEFETs), which convert an electrical signal into an optical signal. In Chapter 4 photodetectors were used, while Chapter 5 deals with LEFETs. For background information about fabrication processes of QD optoelectronic devices see Chapter 2, Section 2.1.3 and 2.5.

62

3.3 Device Fabrication All devices were fabricated on substrates with pre-patterned electrodes. The interdigitated electrodes and the 1.5 x 1.5 mm gold side-electrodes (2 nm Cr, 30 nm Au for PbSe-hybrid photodetectors and 2 nm Ti, 30 nm Au for PbS LEFETs; channel length L = 5 µm; channel width W = 20 mm; W/L ratio = 4000) were defined photo-lithographically using a standard double-layer photoresist (LOR5B/S1813), electron-beam evaporation and a lift-off process on thin glass substrates (Schott AF32 Eco, thickness 0.3 mm) or polyethylene terephthalate (PET) substrates (Figure 3.2). Before the QD material was deposited, the substrates were cleaned with acetone and isopropanol in an ultrasonic bath and subsequently rinsed with deionized water. (a)

(b)

G L S W 8 D Figure 3.2: (a) Schematic illustration of a FET with an interdigitated source-drain electrode structure and a side-gate contact pad. W denotes the channel width and L the channel length. (b) Picture of a glass substrate with three FETs, each with a channel length L of 5 µm. The microscopy image shows a magnification of the channel structure of one FET.

3.3.1 PbSe Quantum Dot Hybrid Photodetectors PbSe-SWNT photodetectors. A PbSe-SWNT hybrid dispersion was drop-cast from hexane on pre-patterned interdigitated electrodes at a temperature of 120 ◦ C, forming the photoactive layer. PbSe-MoS2 photodetectors. A PbSe-MoS2 hybrid dispersion was drop-cast from hexane on pre-patterned interdigitated electrodes at a temperature of 120 ◦ C for glass and 50 ◦ C for PET substrates, forming the photoactive layer (Figure 3.3).

3.3.2 PbS Quantum Dot Light-emitting Field-effect Transistors Fabrication principle. Standard PbS thin-film formation was carried out in a nitrogen glovebox. Different sized PbS QD batches were used, including QDs with

63

3 Experimental Part

Figure 3.3: (a) Schematic illustration of a PbSe-MoS2 photodetector. The MoS2 nanoflakes are shown in gray, the PbSe QDs in green. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.) (b) Picture of a flexible PET substrate with electrode structures for three photodetectors.

diameters of 4.3 nm (absorption maximum λAbs = 1230 nm; PbS 1), 4.6 nm (λAbs = 1303 nm; PbS 2), and 5.1 nm (λAbs = 1407 nm; PbS 3). A solution of PbS QDs (of one size) in octane was passed through a 0.2 µm polytetrafluoroethylene (PTFE) filter (in order to separate PbS aggregates from QDs) onto the sample with prepatterned electrodes. The sample was then spin-coated at 2500 rpm for 10 s to induce a fast evaporation of octane through centrifugal forces and thus ensure a uniform PbS thin-film formation. A 1 % v/v solution of MPA in methanol was dropped onto the PbS QD layer to exchange long oleic acid ligands with short MPA ligands. The MPA solution remained on the PbS layer for 30 s, to allow for a complete ligand exchange, before spinning at 2500 rpm for 10 s. Subsequently, each layer was washed with methanol twice to remove any organic residue and then dried for ∼ 10 s at 80 ◦ C. The complete PbS QD film was formed by a layer-by-layer (LBL) process. In a LBL process the layer deposition and the ligand exchange process are repeated several times until the desired thin-film thickness is reached. PbS QDs outside the interdigitated electrodes were removed with a toluene soaked tissue (see Figure 3.4(a) for PbS QD thin-film formation). Side-gated LEFETs with an ionic liquid dielectric layer. The ionic liquid [EMIM][FAP] was placed inside a polydimethylsiloxane (PDMS) boat on top of the channel structure. A platinum wire, which was partly immersed into the ionic liquid, was used as gate electrode (Figure 3.5(a)). Side-gated LEFETs with an iongel dielectric layer. P(VDF-HFP) was codissolved with [EMIM][FAP] (or [EMIM][TFSI]) in acetone (1:4:14 by mass) and spin-coated on top of the PbS thin-film at 2000 rpm for 1 min. Excess iongel was removed with an acetone soaked tissue. Annealing in nitrogen overnight at ∼ 50 ◦ C removed any residual solvent. All iongel gated devices were encapsulated with a piece of glass and a UV hardening epoxy (Delo Katiobond, LP655 resin) to

64

3.3 Device Fabrication

Figure 3.4: Schematic illustration of a typical PbS QD LEFET fabrication process. (a) PbS QD thin-films were formed on the substrate with a LBL spin-coating process. First, PbS QDs in octane were spin-coated, followed by a ligand exchange with MPA/methanol and a final washing step with methanol. These three steps were repeated for 5-6 times to achieve the desired PbS QD thin-film thickness. (b) The iongel was spin-coated on top of the PbS QD thin-film to form the dielectric layer. (c) For characterization in air, the device was encapsulated with a piece of glass. (d) Picture of one substrate with three PbS QD LEFETs.

allow for measurements in air. The epoxy was hardened by exposing the device with the glued glass plate on top for 3 min to UV radiation (UVA-Cube 100 from H¨onle) (Figure 3.4(b)-(d)). The device was kept in a nitrogen glovebox for another 24 h in order to assure a complete hardening of the epoxy (Figure 3.5(b)). Top-gated LEFETs with a HfOx dielectric layer. 50 nm HfOx were deposited on top of the PbS thin-film by atomic layer deposition (ALD). The sample was first purged 5 min with nitrogen (100 sccm), afterwards it was baked for 20 min under a constant nitrogen flow of 20 sccm. The HfOx layer was deposited using 500 alternating cycles of TDMAH and DI-water at 100 ◦ C. One cycle consists of first, purging 0.015 s with DI-water and second, after a waiting time of 60 s, purging 0.15 s with TDMAH. After another 60 s the next cycle was started. When the dielectric deposition was completed, the device was transferred into a thermal evaporator (Univex 350G) inside a nitrogen glovebox in order to evaporate a 30 nm thick Ag top-gate electrode (evaporation rate: 0.1 A/s up to 5 nm thickness, 0.3 A/s up to 15 nm thickness, 0.5 A/s up to 30 nm thickness) through a metal shadow mask (Figure 3.5(c)). Top-gated LEFETs with a PMMA dielectric layer. The PMMA solution was prepared by mixing electronic grade PMMA (60 mg) with n-butylacetate (1 mL). In order to dissolve the polymer, the solution was heated to 80 ◦ C and kept at this temperature for 2 h under constant stirring. Subsequently, the solution was passed through a 0.45 µm PTFE filter before it was spin-coated on top of the PbS QD thin-film with 2000 rpm for 1 min. Prior to evaporation of a Ag top-gate

65

3 Experimental Part electrode, the device with the PMMA dielectric layer was evacuated overnight at a background pressure of < 106 mbar, ensuring a complete evaporation of n-butylacetate. Finally, the 30 nm thick Ag top-gate electrode (evaporation rate: 0.1 A/s up to 5 nm thickness, 0.3 A/s up to 15 nm thickness, 0.5 A/s up to 30 nm thickness) was evaporated through a metal shadow mask using a thermal evaporator (Univex 350G) (Figure 3.5(d)).

Figure 3.5: Schematic illustration of FETs with different gating strategies. (a) Ionic liquid gated FET with a Pt wire used as gate electrode. A PDMS boat is necessary to keep the ionic liquid on top of the PbS QD thin-film. (b) Iongel gated FET with a side-gate geometry. (c) HfOx and (d) PMMA gated FETs with an evaporated top-gate electrode.

3.4 Material Characterization 3.4.1 Absorption Spectroscopy Absorption spectroscopy is an analytical tool to gather information about the electronic states of lead chalcogenide QDs. If light with a photon energy greater than the band gap of the QDs interacts with dots in the ground state, it will be absorbed, leading to the excitation of an electron from the valence band (VB) into the conduction band (CB), and thus the QD will be transformed from the electronic ground state into the excited state. Due to a certain energy distribution of the electronic states, resulting from the QD size distribution, light absorption will cause a band rather than a single sharp peak in the absorption spectrum. The shape, position, and intensity of the band reflects basic properties of the respective lead chalcogenide QD batch.[511]

Absorption spectra of PbSe-SWNT hybrids and corresponding SWNTs. Absorption spectra were taken with a Varian Cary 5000 absorption spectrometer with a silicon and a PbS detector. Samples were prepared by drop-casting SWNTs or PbSe-SWNTs from hexane onto a glass substrate and subsequently measured under ambient conditions.

66

3.4 Material Characterization Absorption spectra of PbS QDs used for LEFET fabrication. Absorption spectra of PbS QDs were acquired with a Varian Cary 6000i absorption spectrometer with a silicon and an InGaAs detector. PbS QDs were always measured under ambient conditions, either in octane using quartz cuvettes from Hellma Analytics (Suprasil 200, 200-3500 nm; optical path length 1 cm) or after spin-coating on a glass substrate. The wavelength of the first excitonic absorption peak of a QD batch can be deduced from its absorption spectrum. By knowing the wavelength λ, the Planck’s constant h, and the speed of light c, the band gap energy Eg of the respective QD batch can be calculated according to Eg =

hc . λ

(3.1)

Subsequently, the sizes of the PbSe QDs were calculated by solving the following equation:[197] Eg = 0.278 +

0.016d2

1 , + 0.209d + 0.45

(3.2)

where d is the diameter of the PbSe QDs in nm. The sizes of the PbS QDs were determined according to the following equation:[191] Eg = 0.41 +

1 . + 0.283d

0.0252d2

(3.3)

3.4.2 Photoluminescence Spectroscopy The absorption of light transforms the electronic ground state of a QD into an excited states. Photoluminescence (PL) spectroscopy deals with the radiative deactivation of these excited states back into their ground state. For an adequate characterization of PbS QDs and PbS QD thin-films, both, steady-state and timeresolved PL spectroscopy were used. With the help of steady-state PL spectroscopy it was possible to determine the PL quantum yield (PL QY) of PbS QDs in solution and thin-films. The PL QY is the ratio of photons emitted to photons absorbed and it thus provides an indication of PbS QD emission efficiencies. Time-resolved PL spectroscopy allows to monitor PL as a function of time. The PL lifetime is determined by a kinetic analysis of PL decays and relates to the time by which the population of QDs in the excited state, and therefore the PL intensity, has decreased by a certain factor.[94] PL experiments were performed together with Yuriy Zakharko. 67

3 Experimental Part PL spectra. PL characterization of encapsulated PbS QD LEFETs was performed using a custom made setup. PL was generated by exciting PbS QD thin-films within LEFET devices through the glass substrate with a 640 nm laser diode from OBIS (Coherent Europe B.V.). The laser beam was focused onto the LEFETs through a collecting near-infrared (NIR) objective from Olympus (LCPLN50XIR x50, NA 0.65), which exhibits a correction collar in order to take the glass thickness into account. The NIR objective is not only applied to pass the incoming light for PbS excitation, but also to collect emission and pass it to the detector. A cold mirror with a transmission > 875 nm and a long-pass filter (715 nm) between cold mirror and detector were used to reject visible and scattered laser light. In order to record PL spectra an Acton SpectraPro SP2358 spectrometer with a 150 lines/mm grating (Blaze 1200 nm) and a liquid nitrogen cooled one-dimensional InGaAs line camera from PI Acton (OMA V:1024 1.7) were used. The recorded PL spectra were corrected with the ideal spectrum of an Ocean Optics halogen light source (HL-2000). PL QY. In order to determine the PL QY, a 785 nm laser beam was directed through the entrance port of an integrating sphere from LabSphere with a Spectralon coating. PbS QDs were either characterized in octane enclosed in a quartz cuvette or as a thin-film sample prepared by spin-coating. The cuvette or glass sample was mounted on a PTFE sample holder in the center of the sphere. The QY determination was done according to DeMello et al.[512] The laser beam was directed either onto the sample (direct excitation) or on the wall of the integrating sphere (indirect excitation). The scattered laser light and the PL signal were fiber-coupled to an Acton SpectraPro SP2358 spectrometer. In order to correct for re-absorption and re-emission effects within the integrating sphere, the emission spectra obtained from integrating sphere measurements were compared to PL spectra measured outside the sphere.[513] Moreover, all spectra were corrected for the spectral response of the system with a calibrated, stabilized tungsten-halogen lamp (SLS201/M) from Thorlabs. PL lifetime. PbS QD thin-films were excited through the glass substrate with a 785 nm pulsed diode laser from Alphalas GmbH (< 60 ps, 1 MHz or 10 MHz). The laser beam was focused onto the substrate with the help of a x100 NIR 0.8 NA objective from Olympus. The emitted light was collected with the same objective and directed to an InGaAs/InP single-photon avalanche diode from Micro Photon Devices. With the help of a time-correlated single photon counting module

68

3.4 Material Characterization from Picoquant GmbH (Picocharp 300) the distribution of the arrivial times of photons was recorded at a time bin of 32 ps. In order to get the full-width half maximum (FWHM) of the instrument response function, the PL decay of the low band gap diketopyrrolopyrrole copolymer (DPPT-BT) was acquired. From this measurement the FWHM of the instrument response function was estimated to be ∼ 100 ps.

Excitation densities for PL lifetime measurement. In order to define if singleor multiexciton generation occurs within a QD, it is important to know the average number of absorbed photons per QD, defined in our discussion as the excitation density N . N can be expressed by the photon flux per pulse (Nph ) and the absorption cross section of a QD (σ): N = Nph σ,

(3.4)

where the average photon flux is defined as: Nph =

Pav λ 1 . f hc πrS2

(3.5)

Pav displays the average laser power, f the laser repetition rate, λ the excitation wavelength, h the Planck’s constant, c the speed of light, and rS is the radius of the diffraction limited focal spot (accounting for objective NA = 0.8). Therefore, the average photon flux for the two laser powers, namely 630 nW and 15 µW, are calculated to be Nph1 = 0.088 · 1015 cm−2 and Nph2 = 2.096 · 1015 cm−2 .

(3.6)

In order to determine the excitation density, the absorption cross section has to be calculated. This can be done in two different ways. The first method uses the following formula[514] to calculate σ: σ1 =

4π 4 3 |flf |2 nP bS kP bS ( πrQD ), nλ 3

(3.7)

where λ is the photon wavelength in vacuum, n is the refractive index of the environment (∼ 1.5 considering the glass substrate and the iongel), nP bS and kP bS are the refractive index and the extinction coefficient of bulk PbS at wavelength λ, and rQD is the radius of the PbS QDs. The local field correction factor is

69

3 Experimental Part determined as follows: |flf |2 =

(n2P bS − kP2 bS

9n4 ≈ 0.066. + 2n2 )2 + 4n2P bS kP2 bS

(3.8)

Subsequently the absorption cross sections of PbS 1, PbS 2, and PbS 3 are calculated to: σ1(P bS1) = 1.3 · 10−15 cm2 , σ1(P bS2) = 1.59 · 10−15 cm2 , and σ1(P bS3) = 2.17 · 10−15 cm2 . This method to calculate the absorption cross section is usually applied to diluted QD solutions. However, in this study close-packed PbS QD thin-films were characterized, leading to a possible deviation of the local field correction factor and the corresponding absorption cross section. Therefore, the effective absorption cross section of QDs in thin-films was calculated according to:[515] σ2 =

1 − e−αLf ilm , c(QD)Lf ilm

(3.9)

where Lf ilm is the thickness of the QD thin-film, c(QD) is the concentration of the QDs and α is the absorption coefficient. The absorption coefficient is determined according to α=

4πkP bS , λ

(3.10)

leading to an α of 16 · 104 cm−1 . Thus, the second method yields the following absorption cross sections of PbS 1, PbS 2, and PbS 3: σ2(P bS1) = 1.76 · 10−15 cm2 , σ2(P bS2) = 2.15 · 10−15 cm2 , and σ2(P bS3) = 2.93 · 10−15 cm2 . Finally, multiplying the average excitation photon flux per pulse per QD with the absorption cross section, yields an average number of absorbed photons per QD per pulse of less than one (N ≤ 1) for a laser power of 630 nW. 630 nW was used as laser power for all experiments, except of one experiment, where the effect of charge carrier accumulation on excitation power was investigated (see Chapter 5, Section 5.3.3). There, a laser power of 15 µW was chosen, leading to N > 1. PL average lifetime. The PL average lifetime represents the average arrival time of emitted photons at the detector and is defined as:[115] Z τ= 0



P L(t) dt, A0

(3.11)

where A0 is the intensity of the decay curve P L(t) at time zero. Since PL average lifetimes are longer than the instrument response function (IRF), τ can be directly deduced from equation 3.11 without correction for the IRF. 70

3.4 Material Characterization Calculation of PL radiative exciton lifetimes. The radiative lifetimes τr of all three PbS QD batches can be calculated using the respective PL QYs and the measured lifetimes τm . The PL QY is defined as: QY =

1/τr 1/τm

=

τm , τr

(3.12)

leading to the following equation for the radiative lifetime τr : τr =

τm . QY

(3.13)

Normalized Intensity

The average decay lifetimes for all three PbS QDs in solution are measured to be 220 ns for PbS 1, 320 ns for PbS 2, and 130 ns for PbS 3 (Figure 3.6), yielding exciton radiative lifetimes of τr1 = 220 ns/0.12 = 1.83 µs, τr2 = 320 ns/0.11 = 2.91 µs, and τr3 = 130 ns/0.13 = 1 µs. These exciton radiative lifetimes are in the same range as PbS QD exciton lifetimes of previous reports.[206,516] It is assumed that the radiative lifetimes of QD thin-films are the same as the radiative lifetimes in solution. 1

0.1 IRF PbS 1 PbS 2 PbS 3

0.01 100

200

300

400

Time (ns) Figure 3.6: Normalized PL decays of PbS QDs in octane. PL average decay lifetimes are 220 ns for PbS 1, 320 ns for PbS 2, and 130 ns for PbS 3 (λexc = 785 nm). (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

Transient PL analysis. The measured PL decay curves were fitted either to a mono-exponential function or re-convoluted to a tri-exponential function with Symphotime 64 software (Picoquant), which takes the obtained IRF into account. The tri-exponential function can be expressed as: −t/τ 1

P Lmeasured (t) = IRF ⊗ (A1 e

−t/τ 2

+ A2 e

−t/τ

+ A3 e

3

),

(3.14)

where the coefficients A1 , A2 , A3 were re-normalized N1,2,3 =

A1,2,3 , A1 + A2 + A3

(3.15) 71

3 Experimental Part in order to extract relative contribution of the QDs emitting with different lifetimes (N1 + N2 + N3 = 1 (i.e., 100 %)). The weighted residuals of the tri-exponential re-convolution is calculated according to: R(t) =

Decay(t) − F it(t) p , Decay(t)

(3.16)

yielding low residuals for all three different PbS QD batches (Figure 3.7). PbS 1

PbS 2

0 -1 0

20

40

Time (ns)

Weighted Residuals

(d)

Vg = 2.5 V -0.5 V -1.0 V -1.5 V -2.0 V -2.5 V

1 0 -1 0

20

40

Time (ns)

(e) 4

Vg =

2 0 -2 -4

0

20

40

Time (ns)

2.5 V 2.0 V 1.5 V 1.0 V 0.5 V 0.0 V -0.5 V -1.0 V -1.5 V -2.0 V -2.5 V

Residual (counts/s)

2.5 V -0.5 V -1.0 V -1.5 V -2.0 V -2.5 V

(c) Vg = 2.5 V -0.5 V -1.0 V -1.5 V -2.0 V -2.5 V

1 0 -1 0

20

40

Time (ns)

(f) 4

Vg =

2 0 -2 -4

0

20

40

Time (ns)

2.5 V 2.0 V 1.5 V 1.0 V 0.5 V 0.0 V -0.5 V -1.0 V -1.5 V -2.0 V -2.5 V

Weighted Residuals

Vg =

1

Residual (counts/s)

(b)

Weighted Residuals

Residual (counts/s)

(a)

PbS 3

4

Vg =

2 0 -2 -4

0

20

40

Time (ns)

2.5 V 2.0 V 1.5 V 1.0 V 0.5 V 0.0 V -0.5 V -1.0 V -1.5 V -2.0 V -2.5 V

Figure 3.7: Residuals of the mono-exponential fits of non-linear PL transients of Figure 5.19(a),(b),(c) for (a) PbS 1, (b) PbS 2, (c) PbS 3 and weighted residual (see equation 3.16) of the tri-exponential reconvolution fit of the PL decays in Figure 5.18(d),(e),(f) for (d) PbS 1, (e) PbS 2, (f) PbS 3. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

3.4.3 Raman Spectroscopy Raman spectroscopy is a vibrational spectroscopy technique. Raman bands only appear, if the investigated molecule is able to change its polarizability upon excitation. If a laser hits the respective molecule, there will be an interaction between electromagnetic radiation and matter. Supposed the energy of light is not enough to excite an electron from the VB to the CB (see Section 3.4.1), most of the light will be transmitted through the molecule without any effect, however

72

3.4 Material Characterization a small part (factor 10−4 ) will be scattered. Most of the scattered light will be scattered elastically (Rayleigh scattering), thus keeping the initial wavelength. A small fraction of the initial light (factor 10−8 ) will be scattered inelastically. Inelastic scattering either leads to Stokes or Anti-Stokes scattering, depending on the vibrational state of the investigated molecule. If the molecule is in its ground vibrational state, the photon transfers energy to the molecule, thus exciting it to a higher vibrational state while itself losing energy (i.e., redshift). This process causes Stokes scattering. Anti-Stokes scattering occurs if the photon interacts with a molecule, which is already in a higher vibrational state. Then, the photon gains energy from the molecule, leading to a blueshift of the wavelength of the photon. At room temperature most of the molecules are in the ground vibrational state and only a small number of molecules are in higher vibrational states, this is why the appearance of Anti-Stokes scattering is much less likely than Stokes scattering.[511] The Raman spectrometer, which was used for the experiment of this thesis, was only able to detect Stokes scattering. Raman spectra were acquired at room temperature using a Renishaw inVia Reflex Confocal Raman Microscope with a x100 objective (NA 0.85) and an excitation laser wavelength of 785 nm for SWNTs, 532 nm for MoS2 and FLG, and 633 nm for WS2 . Samples were prepared by drop-casting the respective dispersion onto a glass substrate.

3.4.4 Fourier-transform Infrared Spectroscopy Infrared (IR) spectroscopy is a vibrational spectroscopy technique, however unlike to Raman spectroscopy, the investigated material has to be able to change its dipole moment and not its polarizability upon excitation in order to be IR active. If IR light is directed onto the respective material, it will be attenuated when it transmits the material. In transmission mode, the IR radiation, which remains after transmission, will be detected to create the IR spectrum. In Fourier-transform (FT) IR spectroscopy, data is recorded with an interferometer and retranslated to a spectrum with the help of a Fourier transformation. FT IR spectroscopy enables to simultaneously detect all frequencies, thus avoiding a time consuming scan over all wavelengths.[511] FT IR spectra were recorded at room temperature with a NICOLET 6700 FT-IR spectrometer operating in transmission mode. Samples were drop-cast onto a PTFE-IR card.

73

3 Experimental Part

3.4.5 X-ray Photoelectron Spectroscopy X-ray photoelectron spectroscopy (XPS) is a surface sensitive technique to analyze the elemental composition including oxidation state of the surface of a certain material in a non-destructive way. To conduct XPS analysis, X-rays with a certain energy hν are directed onto the material. These X-rays penetrate into the surface of the sample and subsequently eject electrons from the core levels of the surface elements of the investigated material. The depth of analysis in XPS is determined by the attenuation length of the electrons and is around 3-10 nm. The attenuation length of electrons is ∼ 10 % less than the inelastic mean free path of the electrons. The binding energy EB of an electron is elemental specific and is defined as EB = hν − EK − WSp , where hν is the photon energy of the X-ray, WSp is the work function of the spectrometer, and EK represents the kinetic energy of the emitted electron. In order to get the binding energy EB , the kinetic energy EK has to be determined. This is usually done by analyzing the emitted electrons with a spectrometer, e.g., an electron multiplier. Finally, the intensity of the X-ray is displayed versus the binding energy, enabling a clear identification of elements including their oxidation states.[518] XPS was carried out with a Physical Electronics 5600 instrument using an AlKα (hν = 1486.6 eV) X-ray radiation source. Spectra were recorded at room temperature. XPS experiments were performed by Helga Hildebrand.

3.4.6 Scanning Electron Microscopy Scanning electron microscopy (SEM) is an electron microscopy technique, which is usually used to analyze the surface topography and composition of a sample. In order to get SEM images, a convergent electron beam is focused onto the sample with a probe size of 2-10 nm and then scanned in a raster across the region of interest. The smaller the diameter of the electron probe, the better the resolution of the final image. The size of the electron probe decreases with decreasing probe current, however, the current has to be large enough to generate an acceptable signal-to-noise ratio. The electron beam interacts with the sample surface, generating various signals including secondary electrons (exit energies around 2-5 eV) and backscattered electrons (exit energies around 50 eV), which are the two most important signal sources for SEM images. Usually, only secondary electrons are detected to create images, because due to their small exit depth of only a few nanometers they provide the best resolution. For most common SEM microscopes, the sample needs no special preparation, apart from ensuring that

74

3.4 Material Characterization the sample surface is conducting in order to avoid electrostatic charging effects during the characterization process.[519,520] PbSe-MoS2 layers. SEM images were taken with a Hitachi FE-SEM S4800 microscope at 10 kV acceleration voltage. Samples were prepared by drop-casting PbSe-MoS2 from hexane onto a glass substrate with a pre-patterned interdigitated gold structure. Prior to PbSe-MoS2 deposition, the Si/SiO2 substrate was cleaned with acetone and isopropanol in an ultrasonic bath and subsequently rinsed with deionized water. The sample was mounted onto a SEM holder using a silver conductive varnish. SEM images were recorded by Anja Friedrich. PbS QD thin-films. SEM images were acquired with a Zeiss AURIGA microscope at 1 kV acceleration voltage. Samples were prepared in a LBL process from PbS QD octane solutions (as described in 3.3.2) on a Si/SiO2 substrate. Prior to PbS QD deposition, the Si/SiO2 substrate was cleaned with acetone and isopropanol in an ultrasonic bath and subsequently rinsed with deionized water. The sample was fixed onto a sample holder with silver conductive varnish. Side-view SEM samples were achieved by cleaving the Si/SiO2 substrate and mounting it into a special SEM holder by wrapping it with aluminum foil and fixing it with silver conductive varnish, in a way that the top of the sample sticks out of the sample holder. SEM images were recorded by Florentina Gannott.

3.4.7 Transmission Electron Microscopy In transmission electron microscopy (TEM) electrons are emitted from an electron gun and typically accelerated with 80-400 kV, depending on material properties and resolution demands. The electron beam is conducted in vacuum in order to avoid scattering of the electrons by particles in air. The beam is focused through a vertically aligned condenser lens system, which enables a variable illumination aperture and thus a variable illumination area. Subsequently the beam hits the sample, followed by a strong interaction of the beam with the material, leading to elastic and inelastic scattering of the electrons. The sample has to be very thin (usually around a few tens to a few hundreds of nm) in order to be transparent to most of the electrons. The transmitted and forward scattered electrons will be directed through an additional lens system and then projected onto a fluorescent screen, which is usually coupled to a charge-coupled device (CCD) camera. The magnified images will be projected in the image plane, while diffraction patterns

75

3 Experimental Part are imaged in the back focal plane. The three most popular imaging modes in TEM are bright field (BF) TEM imaging, dark-field (DF) TEM imaging, and high-resolution (HR) TEM imaging. In BF-TEM imaging, images are generated only with the direct beam using an objective aperture, which is introduced into the back focal plane and which blanks all electrons scattered to higher angles. The images in DF-TEM imaging are created with a diffracted beam. In HRTEM imaging, all, mostly coherent electron waves are used to form an interference image based on phase contrast, which can yield resolutions below 0.1 nm. In scanning transmission electron microscopy (STEM), the electron beam is focused with the condenser lens system to a diameter of around 0.1 nm and subsequently scanned in a raster across the sample. In high-angle annular dark field (HAADF) STEM the predominantly elastically scattered electrons (Rutherford scattering) will be collected with a ring detector, thus generating mainly mass-thickness contrast. Electron diffraction methods, like selected-area electron diffraction (SAED) and convergent beam electron diffraction (CBED) allow the determination of the crystalline structure and crystal orientations by Bragg-scattered electrons in the back focal plane and thus help to identify the investigated material or material compounds. Additionally, CBED is often used to determine the thickness of crystalline materials. Moreover, it is also possible to operate a TEM as an analytical electron microscope, determining the elemental composition of small regions on the sample with energy-dispersive X-ray spectroscopy (EDX). An EDX detector collects elemental characteristic X-rays, which are generated by electrons interacting with the sample. The combination of the various imaging, electron diffraction, and analytical techniques render a transmission electron microscope a very powerful instrument in material characterization.[519–521] With electron tomography it is possible to form three-dimensional visualizations of samples. In order to obtain these tomographic representations, a tilt series of projections at different tilt angles is acquired. Since the electron beam is fixed, the sample has to be tilted around a single axis using special TEM holders (Figure 3.8). The tilt series acquisition is followed by a series alignment towards one common tilt axis. This enables a tomographic reconstruction using a specific reconstruction algorithm, which finally leads to a three-dimensional visualization. To get the best possible reconstruction, it is necessary to record many images over the largest tilt angle range as possible. However, there are certain challenges that limit the TEM tilt angle range and therefore lead to a missing wedge of information.[522] When the TEM grid is tilted to high angles, first, the electron beam will be shadowed by parts of the TEM grid or the TEM holder, and second, the thickness of lamella-like

76

3.4 Material Characterization samples, which has to be penetrated by the electron beam, increases. Therefore, in most experiments the tilt angle range is limited to ±60◦ , however, special sample holders and sample geometries enable tilting to higher angles.[522,523] Because of various reasons, i.e., geometry of the investigated sample, sample holder, and calibration of the software, it was possible to obtain tilt series up to tilt angles of ±76◦ in the experiments described in Chapter 4. HRTEM and STEM images, SAED patterns, CBED patterns, and EDX spectra were recorded by Benjamin Winter. He also carried out electron tomography experiments.

Figure 3.8: Schematic illustration of the imaging geometry for a TEM tomography experiment. The relevant material (here: green QD) is mounted on a TEM holder, which is tilted by equivalent increments around a single axis perpendicular to the stationary electron beam in order to record a tilt series. Figure adapted from reference.[523]

Sample preparation. For plan-view characterization, PbS and PbSe QDs were drop-cast onto a copper grid with a thin carbon film, whereas hybrids were dropcast onto a lacey carbon copper grid (both from Plano GmbH). To get a side-view of the PbSe-MoS2 sample, the hybrid dispersion was drop-cast on a silicon wafer substrate and covered with a protective carbon layer before preparing cross-sections using a FEI Helios NanoLab 660 focused ion beam (FIB). A thin lamella was cut using the Ga+ ion beam of the FIB. For final polishing the acceleration voltage of the Ga+ beam was reduced from 30 kV down to 2 kV. The lamella was then transferred to an Omniprobe TEM grid using the lift-out technique. Side-view sample preparation by FIB was carried out by Christel Dieker. BF-TEM and HRTEM. BF-TEM and HRTEM images were acquired with a Philips CM300 UltraTWIN microscope, operated at 300 kV acceleration voltage. Aberration-corrected HRTEM images were recorded using an image-side aberrationcorrected Titan3 80-300 microscope operated at 80 kV acceleration voltage to minimize electron-beam induced damage of the sample.

77

3 Experimental Part SAED. SAED patterns of PbSe-SWNT hybrids and of PbSe-MoS2 hybrids were obtained with a Philips CM300 UltraTWIN operated at 300 kV. All remaining SAED patterns were recorded with an image-side aberration-corrected Titan3 80-300 microscope operated at 80 kV acceleration voltage. CBED. CBED patterns to determine the thickness of the MoS2 flakes were recorded using a Philips CM30 operated at 300 kV acceleration voltage. EDX. EDX was performed using an image-side aberration-corrected Titan3 80300 microscope operated at 80 kV acceleration voltage. Samples were prepared by drop-casting PbSe-hybrid solutions onto a 200 mesh lacey carbon copper grid. The Fe, Cu, and C signals in the spectra are generated by the measuring setup. Fe arises from the pole pieces of the objective lens, the Cu peak comes from the TEM grid and the carbon signal is caused by the lacey carbon film of the TEM grid. Si most probably arises from the vacuum grease used during synthesis, whereas O may come from oleic acid or remaining precursor (PbO) (see Figure 4.3, 4.12, 4.16). Electron Tomography. The data for electron tomography was collected by tilting the TEM grid around a single axis perpendicular to the electron beam. To obtain sufficient contrast between the SWNTs and the crystalline PbSe QDs, HAADF STEM was the imaging technique of choice (minimized diffraction contrast, enhanced mass-thickness contrast) for electron tomography. STEM was performed using an image-side aberration-corrected Titan3 80-300 microscope operated at 80 kV acceleration voltage. The ultrathin single-tilt tomography holder from Fischione (model 2020) was used to acquire the tilt series for the three-dimensional reconstructions with tilt angles ranging from -70 ◦ to +74 ◦ for the first experiment (see Figure 4.9(a),(b) and Figure 4.10(a)) and from -66 ◦ to +68 ◦ for the second experiment (see Figure 4.9(c) and Figure 4.10(b)). For both experiments a continuous tilting scheme and a tilt angle increment of 1-2 ◦ were used. The software Xplore3D from FEI was employed to record both automated tilt series. FEI Inspect3D was used to align the images of each tilt series by cross-correlation and to reconstruct the tomogram (SIRT algorithm, 50 iterations).[524] The final three-dimensional visualization of the data set was performed with the VSG Amira ResolveRT software.

78

3.5 Device Characterization

3.5 Device Characterization PbSe-MoS2 photodetectors were characterized in the third part of Chapter 4, while the entire Chapter 5 deals with a detailed characterization of PbS QD LEFETs. For background information on the working principle of QD optoelectronic devices see Chapter 2, Section 2.5.

3.5.1 Photoresponse Measurements PbSe-SWNTs. Photocurrent was recorded at room temperature in a nitrogen glovebox with an Agilent 4155C semiconductor parameter analyzer. A halogen illuminator (MLC-150c from Motic) was used as light source. The measurements were controlled via a Labview program. PbSe-MoS2 . Photocurrent measurements were performed at room temperature in air with an Agilent 4155C semiconductor parameter analyzer. In order to illuminate devices, a broadband halogen light source (HL 2000 from Ocean Optics) with a 1200 nm cut-off filter from Thorlabs was used. The measurements were controlled via a Labview program. Flexible PET substrates were chucked in a homemade substrate holder, where the bending radius could be changed mechanically (Figure 3.9). (a)

(b)

(c)

Figure 3.9: (a) Picture of a bent flexible PbSe-MoS2 photodetector. (b) Similar photodetector fixed in a home-made manual bending aperture in flat configuration and (c) under curvature.

3.5.2 Light-emitting Transistor Characterization Current-voltage characteristics. Current-voltage characteristics of all transistors were carried out with an Agilent 4155C semiconductor parameter analyzer or a Keithley 2612A source meter either inside a nitrogen glovebox or in air for encapsulated transistors. All measurements were controlled via a Labview program

79

3 Experimental Part and recorded at room temperature. In order to monitor charge carrier trapping, forward- and reverse-sweeps were measured for transfer and output characteristics. Charge carrier mobilities. Electron and hole field-effect mobilities for electrolytegated transistors were extracted from the linear regimes of the ambipolar transconductance curves. Charge carrier mobilities were determined as follows: µlin =

∂Id,lin L , ∂Vg W Ci Vd

(3.17)

where W is the channel width, L the channel length, µlin the linear charge carrier mobility, and Ci the capacitance per unit area of the dielectric layer (for derivation of the formula see Experimental Part, Section 2.5.2). Electron filed-effect mobilities between 0.04 and 0.06 cm2 V−1 s−1 , and hole field-effect mobilities between 0.003 and 0.009 cm2 V−1 s−1 were obtained by using a constant capacitance of Ci = 3.4 µFcm−2 for the [EMIM][FAP] iongel. Thiemann et al. determined the capacitance value of 3.4 µFcm−2 for the pure ionic liquid [EMIM][FAP] without the matrix polymer in a plate-plate geometry with platinum electrodes.[525] Therefore, the effective capacitance of the iongel used in the PbS QD LEFETs presented in this work might differ to some extent, this is why mobility values have to be seen as an orientation to roughly classify the device, but not as exact values. Electroluminescence (EL) of LEFETs. EL characterization of the encapsulated PbS QD LEFETs was performed using a custom made setup. The transistors were contacted at source, drain, and gate electrode and controlled with a Keithley 2612A source meter via a Labview program. A NIR objective from Olympus (LCPLN50XIR x50, NA 0.65), which exhibits a correction collar for glass thickness, was used to direct the EL through an Acton SpectraPro SP2358 spectrometer (with a mirror or a grating of 150 lines/mm; blaze 1200 nm) to the detector. A liquid nitrogen cooled one-dimensional InGaAs line camera from PI Acton (OMA V:1024 1.7) was used to obtain EL spectra, while a thermoelectrically cooled two-dimensional InGaAs camera (Xenics XEVA-CL-TE3, resolution 256 x 320 pixels, sensitivity range 800-1600 nm) was chosen to record images or movies of the transistor emission zone. By flipping a mirror it was possible to switch between the one-dimensional line camera and the two-dimensional camera. All recorded spectra were corrected with the ideal spectrum of an Ocean Optics halogen light source (HL-2000).

80

3.5 Device Characterization External quantum efficiency (EQE). In order to quantify light output of PbS QD LEFETs, EQEs were determined. The transistors were contacted at source, drain, and gate electrode and controlled with an Agilent 4155C semiconductor parameter analyzer via a Labview program. For current-voltage luminance measurements, a calibrated InGaAs photodiode (Thorlabs FGA21-CAL; active area 3.1 mm2 ) with a detection range between 800 and 1800 nm was placed directly underneath the channel area of the respective transistor (active area ∼ 0.1 mm2 ), which was fabricated onto a transparent glass substrate. However, there are some expected losses that might underestimate EQEs. First, light emission happens randomly in all directions and the photodiode is only detecting the light emitted towards the photodiode, second there might be some absorption and waveguiding losses of the emitted light between the emission source (PbS QDs) and the detector, and third, for PbS 2 and PbS 3 it is not possible to detect the whole emission band, due to the limited detection range of the photodiode. Therefore, the obtained EQEs should be considered as lower boundaries of the overall efficiency. In order to be able to calculate EQEs, the transistors were biased as for usual transconductance measurements, i.e., with a constant drain voltage and a varying gate voltage. The photodiode was set to short-circuit conditions, which means 0 V across cathode and anode. Then, the photocurrent Iph was measured, enabling the calculation of the EQE by taking into account the EL spectrum and the wavelength dependent sensitivity of the photodiode. In detail, EQE calculation was done as follows:[526] The photocurrent Iph is a product of the output power P and the sensitivity of the photodiode S. Normally, the sensitivity of a photodiode is not constant across its whole wavelength range, therefore the sensitivity of the diode has to be averaged by weighing with the normalized EL spectrum ELnorm (λ), resulting in an adapted sensitivity Save : R Save =

S(λ)ELnorm (λ) dλ R . ELnorm (λ) dλ

(3.18)

This leads to: P =

Iph . Save

(3.19)

Subsequently, the number of emitted photons Nem per second can be deduced from the output power P and the photon energy E by the ratio P/E with E = hc/λ, where h is the Planck constant, c the speed of light, and λ the wavelength. In order to calculate the total number Nep of emitted photons per second, the P/E ratio

81

3 Experimental Part still hast to be weighted with ELnorm (λ) and integrated over the whole wavelength range, leading to: Nep

R P λELnorm (λ) dλ R . = hc ELnorm (λ) dλ

(3.20)

Then, the EQE is calculated by dividing the total number of photons Nep through the ratio of drain current Id to elementary charge e: EQE =

Nep . Id/e

(3.21)

Thus, by inserting equation 3.18 - 3.20 in equation 3.21 the EQE of the PbS QD LEFETs can be simplified to: R R λELnorm (λ) dλ Iph 1 Iph 1 e λELnorm (λ) dλ −4 R = . EQE = ·8.066 · 10 · R Id Save hc ELnorm (λ) dλ Id Save ELnorm (λ) dλ (3.22) Finally, in order to display EQE depending on the current density, the current density [A/cm2 ] was calculated, taking into account the total channel width (20 mm), the height of the PbS QD thin-film (∼ 220 nm, determined from SEM), and the source-drain current.

82

4 Colloidal PbSe Quantum Dot Hybrid Materials

The following chapter discusses a new and facile method to synthesize PbSe quantum dot nanohybrids and the fabrication of near-infrared sensitive photodetectors with these hybrids. In a one-step hot-injection synthesis, PbSe quantum dots are directly and non-covalently grown on untreated single-walled carbon nanotubes, few-layer graphene and transition metal dichalcogenides. PbSe quantum dots form half-rings around single-walled carbon nanotube bundles, while they grow epitaxially on MoS2 nanoflakes. Finally, PbSe-MoS2 hybrids were used to fabricate air-stable nearinfrared sensitive photodetectors on rigid glass and flexible polymer substrates.

4 Colloidal PbSe Quantum Dot Hybrid Materials

4.1 Introduction Semiconductor quantum dots (QDs) exhibit many promising properties, for example they are relatively easy to synthesize, they are solution-processable, and most important they exhibit size-tunable optical properties (see Chapter 2, Section 2.1). In order to fabricate well performing QD-based devices, an efficient charge transport within QD thin-films is essential, however, this is difficult to achieve.[116,117] One possible solution is to couple QDs to high-mobility nanomaterials in order to form nanohybrids. These hybrids have to fulfill some requirements to be interesting for potential applications, i.e., they should keep the remarkable properties of QDs, which are wet-chemical processing and size-tunable absorption. Additionally they should provide an efficient charge transfer from the QDs to the supporting material with a subsequent fast charge transport in the high-mobility material, in order to achieve photodetectors with high photosensitivity and short response times. This chapter deals with a new method to synthesize PbSe QD nanohybrids, which are then used to fabricate near-infrared (NIR) sensitive photodetectors. In a simple one-step hot-injection synthesis, PbSe QDs are directly and noncovalently attached to a wide range of one- and two-dimensional materials, such as untreated single-walled carbon nanotubes (SWNTs), few-layered graphene (FLG), and transition metal dichalcogenides (TMDs). At first, investigations concentrate on PbSe-SWNT hybrids. By varying the synthesis conditions, the size- and shapedependence of PbSe QDs in in situ grown PbSe-SWNT hybrids is investigated. The application of high-resolution transmission electron microscopy (HRTEM) and electron tomography reveals the three-dimensional morphology and atomic structure of the PbSe-SWNT hybrids. In a next step, the synthesis developed for PbSe-SWNT hybrids is transferred to PbSe-layered material hybrids, replacing carbon nanotubes (CNTs) with graphene and/or TMDs. PbSe-FLG, PbSe-WS2 , and PbSe-MoS2 are synthesized and characterized, whereby the interface between QDs and layered materials is investigated in detail for PbSe-MoS2 hybrids using selected-area electron diffraction (SAED) and cross-sectional transmission electron microscopy (TEM). Finally, the PbSe-MoS2 hybrids are used to fabricate NIR sensitive photodetectors on rigid glass and flexible polyethylene terephthalate (PET) substrates. Most of the TEM experiments were performed in cooperation with Benjamin Winter.

84

4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids

4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids 4.2.1 Fabrication of PbSe-SWNT Hybrids PbSe-SWNT hybrids were synthesized using the hot-injection synthesis technique. At 120 ◦ C a dispersion of SWNTs in diphenyl ether was injected into a stocksolution of Pb-oleate, followed by a temperature increase to 170 ◦ C. Then, a Se-precursor solution was quickly injected starting the QD growth. After a certain growth time the reaction was quenched and PbSe-SWNT hybrids were isolated using the nonpolar solvent hexane. Since hexane dissolves PbSe QDs, while the hybrid precipitates in hexane, a facile separation of both products is achieved (for details on synthesis see Experimental Part, Section 3.2). Figure 4.1 depicts bright field transmission electron microscopy (BF-TEM) images of PbSe-SWNT hybrids, indicating that PbSe QDs are attached to a network of SWNT bundles. It becomes clear that there is no higher order or orientation of PbSe QDs on SWNTs, it rather seems that QDs are somehow randomly distributed across the SWNT network. Moreover, the BF-TEM images show no unbound QDs, excluding a simple arrangement of PbSe QDs onto SWNT sidewalls due to drying effects resulting from a solvent evaporation on the TEM grid. (a)

200 nm

(b)

200 nm

(c)

40 nm

Figure 4.1: (a) and (b) Typical BF-TEM images of PbSe-SWNT hybrids from two different sample positions and (c) BF-TEM image with a higher magnification, showing that PbSe QDs are attached to SWNT bundles rather than single nanotubes (Pb:Se ratio 1:1; growth time 5 min). (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

Depending on the further application of the CNTs, the exfoliation solvent of the SWNTs has to be chosen carefully by the ability to fit in the overall synthesis process. The hot-injection PbSe QD synthesis would be disturbed by solvents like sodium dodecyl sulfate (SDS) or n-methyl-2-pyrrolidone (NMP), therefore these reasonable good CNT exfoliation solvents could not be utilized for the here presented hybrid synthesis. Moreover, perylenes and polymers also exfoliate

85

4 Colloidal PbSe Quantum Dot Hybrid Materials and subsequently stabilize exfoliated CNTs, however these molecules cover the SWNT surface and thus would prevent a direct contact between QDs and SWNTs. Diphenyl ether is able to partly realize nanotube debundling while still enabling a QD hot-injection synthesis and thus was used for all following syntheses. Therefore, SWNTs always appear as bundles and not as separated tubes.

4.2.2 Characterization of PbSe-SWNT Hybrids In order to investigate the PbSe QD shape and the connection between QDs and SWNT sidewalls in more detail, HRTEM characterizations were conducted. All HRTEM images were recorded with an acceleration voltage of 80 kV in order to keep electron beam induced damages of the SWNTs as small as possible. In Figure 4.2 it can clearly be seen that there is an intimate contact between the two materials, which renders the presence of capping ligands at the interface very unlikely. For this reason, it is assumed that SWNT bundles act as ligands for the PbSe QDs, which also implies the absence of capping ligands at the interface. This theory is supported by a similar statement in previous work by Ju´arez et al.,[380,381] discussing cadmium chalcogenide QDs on CNTs. (a)

(b)

(c)

0.31 nm

5 nm

5 nm

5 nm

Figure 4.2: HRTEM images of PbSe-SWNT hybrids, illustrating the intimate contact between PbSe QDs and SWNT bundles, and the preferred orientation of the PbSe QDs with their {002} lattice planes (d = 0.31 nm; depicted in yellow in image (a)) perpendicular to the SWNT bundles. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

Although TEM images are very helpful to demonstratively show the PbSeSWNT nanohybrid, they provide no information about the chemical composition of this material. Therefore, the nanohybrid was analyzed by energy-dispersive X-ray spectroscopy (EDX), indicating lead, selenium, and carbon as the primarily present elements (Figure 4.3). There are also some other peaks, e.g., Cu, and Fe, however, these peaks arise from the measuring setup (see Experimental Part, Section 3.4.7 for details) and do not indicate impurities in the synthesized PbSe-SWNT hybrid material. 86

4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids

Counts (x 103)

8

C

Si

Pb Se

O

4 Cu

Co/Fe Cu

Fe

0

0

4

8

12

16

Energy (keV) Figure 4.3: EDX spectrum of PbSe-SWNT hybrids showing mainly lead (“red”) and selenium (“green”) peaks. Cu, Fe, Co, Si, O, and (partly) C peaks can be ascribed to residuals from synthesis and elements arising from the measuring setup. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

HRTEM images clearly show the crystal lattice planes and thus the single crystalline nature of PbSe QDs without extended defects (Figure 4.2). It is striking that the {002} lattice planes with a lattice constant of 0.31 nm of most of the PbSe QDs are oriented perpendicular to the SWNT bundles. SAED further confirms the rock-salt crystal structure of the PbSe QDs (Figure 4.4(a)). A comparison between the experimentally measured rotationally-averaged diffraction pattern and the theoretically expected peak intensities of PbSe-SWNT nanohybrids shows some deviations, most striking a much higher {002} reflection peak intensity in the experimental data (Figure 4.4(b)). There are some factors that might generate small differences between the two spectra, e.g., background by the SWNTs, the TEM carbon grid, and possible influences by multiple scattering. However, the main reason of this deviation is most likely the already observed preferred orientation of the {002} PbSe lattice planes perpendicular to the SWNT bundles. Looking more closely at the BF-TEM and HRTEM images, the preferred orientation of the PbSe QDs always appears where PbSe QDs were able to grow on one single thin SWNT bundle without any contact to one or more other SWNT bundles. This preferential orientation of PbSe QDs in relation to SWNT bundles was also reported by Ka et al.,[374] where PbSe-SWNT hybrids were fabricated via a pulsed laser ablation technique. While the interface between PbSe QDs and SWNTs was already discussed, the surface of the QDs toward the solvent has not yet been investigated. Fourier transform infrared (FT IR) spectra show aliphatic stretching modes at 2848 cm−1 and 2917 cm−1 (Figure 4.5(a)). These signals can be assigned to alkyl groups and thus confirm the presence of stabilizing ligands somewhere within the PbSe-SWNT nanohybrid. Oleic acid and tri-n-octylphosphine (TOP) are the only two possible

87

0.0

3

4

5

6

{024} {133}

0.5

{004}

1.0

{222}

(b)

{002}

8 579 46

1: {111} 2: {002} 3: {022} 4: {113} 5: {222} 6: {004} 7: {133} 8: {024} 9: {224}

{113}

1

3

{111}

2

Intensity

(a)

{022}

4 Colloidal PbSe Quantum Dot Hybrid Materials

7

-1

g (nm )

Figure 4.4: (a) SAED pattern of PbSe-SWNT hybrids, confirming the rock-salt crystal structure of PbSe QDs. The yellow peaks indicate the expected rotationally-averaged peak intensities. (b) Experimentally recorded rotationally-averaged diffraction pattern of PbSeSWNT hybrids (black line) and carbon (SWNTs and carbon grid) (blue dotted line), after background correction. The theoretically expected peak intensities are shown with red bars. In the experimental data the {002} reflection peak is much higher than theoretically expected, indicating the preferred orientation of the {002} PbSe lattice planes perpendicular to the SWNT bundles. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

surfactants used during the synthesis process, which also contain alkyl groups. Therefore, either oleic acid or TOP has to be the ligand from the PbSe QDs towards the solvent. X-ray photoelectron spectroscopy (XPS) shows no phosphorous peak, this is why TOP was excluded (Figure 4.5(b)) and oleic acid was identified as the present stabilizing ligand. By knowing the ligand, it is possible to draw a conclusion on the surface composition of the PbSe QDs within the hybrid structure and therefore to get a better understanding of the underlying crystallographic structure. Oleic acid stabilizes Pb-terminated QDs, whereas TOP would bind to a Se-terminated surfaces. Thus, FT IR in combination with XPS confirms a Pb-terminated QD surface. This result is in line with previous reports on PbSe QDs, where the surface is usually Pb-terminated (see Chapter 2, Section 2.1.4). Consequently, the in situ growth of PbSe QDs in presence of SWNTs seems to have no influence on the usual surface termination of PbSe QDs. Not only a direct, ligand-free interface between QDs and SWNTs is important for a possible application of the nanohybrid in devices, but also a non-covalent binding of the QDs to the SWNTs is essential. Through non-covalent binding the SWNT carbon lattice remains intact and thus the high charge carrier mobility of the SWNTs is maintained. Raman spectroscopy represents an informative technique to study the chemical bonds between quantum dots and carbon nanotubes. SWNTs exhibit certain characteristic Raman bands, which were introduced in Chapter 2, Section 2.2.2. The intensity of the D-band, which appears at around 1290 cm−1 , quantifies the amount of sp3 -hybridized carbon atoms and thus the defect density in

88

4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids (b)

100

12 3

Intensity (x 10 )

Transmission (%)

(a)

CO2

80 -1

2848 cm -1 2917 cm

3200

2400

1600 -1

Wavenumber (cm )

Pb4f5

Pb4f7

8 4 0 148

P2p

144

140

136

132

Binding Energy (eV)

Figure 4.5: (a) FT IR of SWNTs (“purple”) and PbSe-SWNT hybrids (“green”). The hybrid spectrum shows aliphatic stretching modes at 2848 cm−1 and 2917 cm−1 , indicating alkyl groups in the sample. (b) XPS spectrum of PbSe-SWNT hybrids, demonstrating the absence of phosphorous in the sample. The P2p peak would appear at around 134 eV. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

a SWNT batch. Figure 4.6(a) shows, that the intensities of the D-bands are similar for pure SWNTs and for PbSe-SWNT hybrids. This implies that the growth of PbSe QDs around SWNTs does not significantly disturb the sp2 -hybridized carbon lattice of the SWNTs, leading to the conclusion that PbSe QDs are non-covalently attached to the SWNT bundles. Besides Raman spectroscopy, absorption spectroscopy additionally confirms a non-covalent bonding. The absorption spectra of pure SWNTs and PbSe-SWNT hybrids show the same two absorption bands (Figure 4.6(b)). The first one appears between 900 and 1500 nm, corresponding to the E11 exciton transition, while the second one between 450 and 800 nm arises from the E22 exciton transition. The fact that both spectra resemble each other indicates that the electronic structure of the SWNTs is maintained throughout the hybrid synthesis process, which supports a non-covalent bonding between QDs and SWNTs. A covalent attachment of the PbSe QDs to the SWNTs would destroy the SWNT electronic structure and consequently result in a reduction of the absorption bands up to a complete disappearance of both features. Moreover, the absorption spectrum of PbSe-SWNT hybrids shows no absorption bands corresponding to the PbSe QDs. Possible explanations might be the polydispersity of the QDs leading to a very broad and therefore hardly observable QD peak and/or the strong SWNT background absorption overlaying the PbSe QD absorption feature. However, despite of the corroborated non-covalent attachment, the bonding between QDs and SWNTs is quite strong. Even ultrasonication is not able to remove the QDs from the SWNT surface.

89

4 Colloidal PbSe Quantum Dot Hybrid Materials (a)

(b)

Intensity

0.1

D

0.8 1200

1300

1400

0.4 0.0

0

1000

0.4

Absorbance

1.2

2000 -1

Raman Shift (cm )

E22 E11 0.2

0.0

1000

2000

Wavelength (nm)

Figure 4.6: Raman and absorption spectra of SWNTs (“purple”) and PbSe-SWNT hybrids (“green”). (a) Raman spectra (normalized to G peak at 1587 cm−1 ) show, that the D-band intensity is the same for both materials, indicating that the carbon lattice is not damaged by the hybrid synthesis (λexc = 785 nm) (b) The same shape of both absorption spectra indicates that the electronic structure of the SWNTs is maintained throughout the synthesis process. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

To summarize this first part it can be stated that it is possible to synthesize non-covalently bonded PbSe QD-SWNT hybrids in a one step in situ synthesis. However, regarding future applications it is not only important to produce the hybrid but also to have good control over the size and shape, which will be discussed in the following section.

4.2.3 PbSe QD Size and Shape Optimization In order to precisely tune size and shape of the hybrid, a synthesis series with different growth times and a second synthesis series with different Pb to Se ratios was carried out to identify the influence of the varying growth conditions. Within one series, only one parameter, either growth time or Pb to Se ratio, was changed, while the second parameter was kept constant. For the first synthesis series, different growth times where chosen, namely 60 s, 5 min, and 24 h, while the Pb to Se ratio was kept constant at 1:1. Figure 4.7 shows BF-TEM images of the PbSe-SWNT hybrids after all three growth times. As expected from a hot-injection QD synthesis, the PbSe QD size increases with increasing growth time. After 60 s growth time the synthesis solution exhibits a brownish color. Unbound PbSe QDs are responsible for the colored solution (see inset Figure 4.7(a)), indicating many unbound PbSe QDs besides the PbSe-SWNT hybrids. Even though the hybrid can be separated from the unbound QDs during the purification process, PbSe-SWNT hybrids exhibit a low PbSe QD load after 60 s growth time. The average size of these PbSe QDs ranges between 5-6 nm (Figure 4.7(a)). A growth time of 5 min yields a colorless solution, which means that the

90

4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids PbSe QDs are almost entirely attached to the SWNTs, leaving no unbound PbSe QDs in the final synthesis solution. After 5 min growth time the hybrid exhibits QDs with 8 to 10 nm in size (Figure 4.7(b)). BF-TEM images show, that PbSe QDs grow anisotropically. The initially spherical-shaped QDs start to get more faceted and simultaneously seem to grow around the SWNTs. After 24 h growth time the synthesis solution is dark black with a metallic luster. This observation is a sign of a partial decomposition of PbSe into elementary lead. However, BF-TEM images indicate, that still some PbSe-SWNT hybrids have been grown (Figure 4.7(c)). PbSe QDs are faceted and around 20-35 nm in size. They partly seem to form complete PbSe rings, enclosing the SWNT bundles. (a)

5 nm

(b)

60 s

5 nm

(c)

5 min

10 nm

24 h

Figure 4.7: BF-TEM images of PbSe-SWNT hybrids (Pb:Se ratio = 1:1) after (a) 60 s, (b) 5 min, and (c) 24 h growth time. The inset in (a) shows an unbound PbSe QD, which was separated from the hybrid; scale bar inset = 5 nm. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

For the second synthesis series the initial Pb to Se ratio was changed from a shortfall of Pb (Pb:Se = 1:2) to an equal ratio of Pb to Se (Pb:Se = 1:1) up to an excess of Pb (Pb:Se = 2:1) (Figure 4.8). The growth time was set to 5 min for all experiments. First, PbSe QDs with all three Pb to Se ratios were synthesized without the presence of SWNTs. However, in consideration of comparability, diphenyl ether was nevertheless injected into the lead oleate solution before adding the selenium precursor (see insets in Figure 4.8). For a 1:2 Pb to Se ratio the hot-injection synthesis yields spherical and monodisperse QDs. With increasing Pb ratio, the QD shape changes from spherical to more faceted and the QDs in one batch exhibit a larger size distribution. Therefore, in order to synthesize a QD batch with high monodispersity, a Pb to Se ratio of 1:2 should be chosen. In a next step, it was investigated if the Pb to Se ratio, which appears to be perfect to synthesize spherical and monodisperse QDs, also yields the best results for the PbSe-SWNT hybrids. Figure 4.8(a) shows that for a Pb to Se ratio of 1:2, PbSe QDs seem to settle on the SWNTs, whereas for a 1:1 ratio (Figure

91

4 Colloidal PbSe Quantum Dot Hybrid Materials 4.8(b)) QDs seem to grow around the SWNT bundles. A further increase of the Pb ratio to Pb:Se 2:1 results in an even better wetting of the SWNT sidewalls by the QDs (Figure 4.8(c)), however, the PbSe load is a lot less than for the 1:1 ratio. Therefore, the 1:1 ratio was chosen as the optimal ratio to synthesize PbSe-SWNT hybrids. Interestingly, by comparing these results with literature reports, a Pb to Se ratio of 1:1 lies in between the ideal ratio for a spherical PbSe QD formation (Se excess is necessary) and the perfect ratio for a PbSe nanowire formation (Pb excess is necessary).[61] The two presented synthesis series clearly show that it is possible to control the size and shape of the PbSe-SWNT nanohybrids. For all following experiments the previously determined ideal parameters were used, namely a growth time of 5 min and a Pb to Se ratio of 1:1. (a)

20 nm

(b)

1:2

20 nm

(c)

1:1

10 nm

2:1

Figure 4.8: BF-TEM images of PbSe-SWNT hybrids (growth time = 5 min) for a Pb:Se ratio of (a) 1:2, (b) 1:1, and (c) 2:1. The insets show the corresponding QDs, grown under the same conditions as the hybrids, only without SWNTs; scale bar insets = 10 nm. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

4.2.4 Three-dimensional Morphology The optoelectronic properties of PbSe-SWNT nanohybrids strongly depend on the interface between the two materials and the shape and size of the QDs. In Section 4.2.2 the absence of linker molecules at the interface between PbSe QDs and SWNTs was corroborated by different spectroscopic characterization methods while the size and shape of the QDs was examined by TEM in Section 4.2.3. However, TEM is a two-dimensional imaging technique, therefore it is only possible to assume that the QDs grow around the SWNTs using the bundles as a template, but so far no direct evidence was provided. In order to address this fact, electron tomography was chosen to characterize the hybrid structure in three dimensions. In an electron tomography experiment the TEM grid is tilted around a single axis perpendicular to the electron beam. Two single-axis tilt series of different, but equally synthesized PbSe-SWNT batches were conducted, where in the first experiment the TEM grid was tilted from -70◦ to +74◦ , while in the 92

4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids second experiment it was tilted within a slightly smaller angle range from -66◦ to +68◦ (Figure 4.9). Two-dimensional overview images were recorded with scanning transmission electron microscopy (STEM) (exemplary shown for tilt series one in Figure 4.9(a)). Subsequently, for batch one, one representative particle was chosen to perform the tilt series. This particle is encircled in yellow in Figure 4.9(a). For the second tilt series two adjacent particles were selected. In order to get a tilt series, roughly one hundred STEM images (103 STEM images for batch one, 95 STEM images for batch two) were recorded, all at different tilt angles. Figure 4.9(b) shows three representative STEM images of tilt series one, while Figure 4.9(c) shows three representative STEM images of the second tilt series. SWNTs are barely visible, while PbSe QDs appear much brighter. This observation results from the enhanced mass-thickness contrast in STEM imaging mode and the relatively large difference between the atomic number of carbon and PbSe. The three-dimensional morphology of the recorded PbSe-SWNT hybrid structure was finally reconstructed using all STEM images of one tilt series. The morphologies of both batches clearly show that the QDs not only sit on the SWNT sidewalls but rather grow around the SWNTs using the bundles as a template within the growth process (Figure 4.10). However, not only the shape but also the correlation between shape and crystal orientation of the PbSe QDs toward the SWNTs can provide useful information enabling inference of the growth mechanism. It was possible to gain this information, because the PbSe QDs exhibit a preferred crystal orientation relating to the long axis of SWNT bundles ({002}P bSe perpendicular to the SWNT bundle) and the tilt axis was chosen parallel to the SWNT bundle. For batch number one, Figure 4.11(a) shows projection views (surface rendering) of the three-dimensional reconstructed PbSe QD along the [001] and [011] axis. By comparing the HRTEM images (Figure 4.11(b)) with the reconstructed PbSe QD shape at the same tilt angle, it can be seen that both shapes coincide very well. Figure 4.11(c) depicts the projection view (surface rendering) of the three-dimensional reconstructed QD along the tilt axis. The green and yellow arrows indicate the orientation of the [001] and [011] zone axis perpendicular to the axis.

4.2.5 Growth Mechanism Considering all experimental results and combining them with previous literature reports, it is possible to propose a growth mechanism for the in situ PbSe-SWNT hybrid formation.[59,61,195,528,529] Based on the information gained from changing growth time and Pb to Se ratio, it can be stated that depending on the initial Pb 93

4 Colloidal PbSe Quantum Dot Hybrid Materials TILT SERIES 1 (BATCH 1) (a)

1 µm

(b)

100 nm

a = +40°

a = 0°

20 nm

20 nm

(c)

TILT SERIES 2 (BATCH 2) a=0

50 nm

a = -40°

50 nm

a = -52°

20 nm

a = +40°

50 nm

Figure 4.9: (a) Overview STEM images and (b) three representative STEM images (out of 103 images) of tilt series one at different tilt angles (+40◦ , 0◦ , -52◦ ). The investigated particle is encircled in yellow. (c) Three representative STEM images (out of 95 images) of tilt series two at different tilt angles (-40◦ ; 0◦ ; +40◦ ). In tilt series two, two adjacent particles were analyzed. α indicates the tilt angle. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

to Se ratio, PbSe QDs form both, on SWNT sidewalls and in solution. However, for the finally used synthesis conditions (5 min growth time and a Pb to Se ratio of 1:1), PbSe QDs are mainly found to be attached to SWNT bundles. At this point it is assumed that the metal precursor (Pb-oleate) coordinates onto the surface of the SWNT bundles and thus initiates the formation of the hybrid. The nucleation mechanism was investigated in more detail in order to strengthen this assumption. This was done by the isolation of a Pb-SWNT precursor before injection of TOPSe. After several purification cycles the Pb-SWNT precursor was characterized via 94

4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids (a)

(b)

5 nm

5 nm

Figure 4.10: Reconstructed volume of (a) the single QD investigated during tilt series one and of (b) the two adjacent QDs characterized during tilt series two, showing that the QDs grow around the SWNT bundles. Silver-yellow surface rendering indicates lead selenide, while green-blue volume rendering indicates carbon. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

(a)

(b)

(c) [100] [011]

10 nm

10 nm

[001]

10 nm

Figure 4.11: (a) Projection views (surface renderings) and (b) corresponding HRTEM images for the QD of tilt series one, clearly showing the coinciding shapes. The QDs are tilted into different zone axes (upper image view along [011], -tilt 24.5◦ ; lower image view along [001], -tilt +20◦ ; tilt axis horizontally positioned within image plane). (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

EDX. The EDX spectrum in Figure 4.12 reveals a high concentration of Pb in the Pb-SWNT sample before TOPSe injection, and thus supports the assumption of a metal precursor coordination onto the surface of the SWNTs. There are different possible explanations for the nucleation of lead onto the SWNT sidewalls, which might coexist to some part. First, SWNTs exhibit a large extended electronic π-system, which is able to compensate charges. Therefore, SWNTs might attract and subsequently stabilize Pb2+ cations. Second, even non-pretreated SWNTs exhibit a certain amount of surface defects (see D-band in Raman spectroscopy), which can act as nucleation centers for lead cations. The third probable explanation is that Co/Mo catalyst particles, which remained from

95

4 Colloidal PbSe Quantum Dot Hybrid Materials 1

Counts (x103)

C Si O

Pb

Cu

Cu

Co/Fe

Cu

Fe

0

0

4

8

12

16

Energy (keV) Figure 4.12: EDX spectrum of Pb-SWNT hybrids before TOPSe injection and after several washing cycles, showing a high lead (“red”) concentration. Cu, Fe, Co, Si, O, and (partly) C peaks can be ascribed to residuals from synthesis and elements arising from the measuring setup. (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

SWNT production, can act as nucleation sites for Pb2+ cations (see Chapter 2, Section 2.2.1). However, STEM images of pure SWNTs show that the amount of Co/Mo particles in the used SWNT batch is a lot less than the final PbSe QD concentration on the SWNTs (Figure 4.13). Therefore, these catalyst particles might contribute to PbSe nucleation but cannot be the only explanation. (a)

(b)

100 nm

100 nm

Figure 4.13: (a) STEM image of CoMoCAT-SWNTs as received, dispersed in diphenyl ether. The bright spots indicate the presence of impurities, including Co and Mo catalyst particles. (b) STEM image of PbSe-SWNT hybrids. Comparing both images clearly indicates a substantially higher concentration of PbSe QDs in (b) than catalyst particles in (a). (Figure adapted from Schornbaum et al.[527] Copyright 2013 American Chemical Society.)

The next step of the growth process after the nucleation of lead is initiated by the injection of a selenium precursor, which starts the in situ growth of the PbSe QDs. The reconstructed three-dimensional morphology of the PbSe QDs clearly shows that the SWNT bundles act as templates for the QD growth. After an initial formation of small PbSe QDs, the dots might further grow via oriented attachment. PbSe QDs exhibit a dipole moment along the h001i axis (see Chapter 2, Section 2.1.4 for the explanation of a dipole moment in a centrosymmetric rock-salt crystal),[61,195] which can induce a spontaneous oriented attachment of

96

4.2 PbSe Quantum Dot-Carbon Nanotube Hybrids small QDs along the h100i direction. If the growth time is extended (up to 24 h), the reaction shifts from the thermodynamic growth regime to the kinetic growth regime (see Chapter 2, Section 2.1.1 for further explanation), forming faceted and almost cubic PbSe QDs. Moreover, the concentration of unbound QDs in solution decreases with longer growth times, while simultaneously the PbSe QDs, which are already attached to SWNT bundles, continue to grow. This can be explained by a favored oriented attachment of QDs, which are bound to SWNTs, over oriented attachment in solution and is in line with published results by Talapin et al.,[530] where a screening of the electric field of PbSe dipoles by the free electrons of a carbon π-system (e.g., SWNTs) was reported. It can be concluded that the PbSe QDs, which grow on the SWNT sidewalls, are stabilized by the carbon π-system of the SWNTs. Therefore, the growth of PbSe QDs on SWNT bundles is thermodynamically favored.

4.2.6 PbSe-SWNT Photodetectors

Current (µA)

The aim of synthesizing PbSe-SWNT hybrids was their use in solution-processable, NIR photodetectors. For this reason, photodetectors were fabricated by dropcasting a solution of ultrasonically dispersed PbSe-SWNT hybrids on a glass substrate with pre-patterned interdigitated gold electrodes. SWNTs always come as a mixture of semiconducting and metallic CNTs and within the scope of the experiments it was not possible to separate metallic from semiconducting nanotubes. Therefore, photodetectors made of PbSe-SWNTs already yield very high dark currents. Due to the relatively low concentration of PbSe QDs compared to SWNTs within the PbSe-SWNT hybrids, it is impossible to detect a photocurrent, which reliably results from a PbSe QD photosensitization of the SWNT bundles (Figure 4.14). light on light off

2 0 -2 -1

0

1

Bias (V) Figure 4.14: Current vs. bias voltage of a PbSe-SWNT photodetector. The black line displays the dark current, while the red line indicates the current under illumination.

97

4 Colloidal PbSe Quantum Dot Hybrid Materials

4.2.7 Summary A new synthesis route was developed to effectively grow PbSe QDs on SWNT bundles, forming PbSe-SWNT nanohybrids. With this new synthesis technique it is not only possible to achieve PbSe-SWNT nanohybrids but also to tune the size, shape, and concentration of the assembled QDs on the SWNTs. An in situ synthesis route was chosen in order to guarantee a good electronic coupling between the two materials. The presence of linker molecules between PbSe QDs and SWNTs is avoided and consequently a direct interface in between both materials is provided. Moreover, the synthesis was performed without chemical modification of the SWNTs, hence it was possible to maintain the excellent charge transport properties of the one-dimensional nanotubes. The non-covalent coupling between the two materials was corroborated by Raman and absorption spectroscopy. HRTEM images clearly show that the PbSe QDs with a rock-salt crystal structure are intimately attached to the SWNT bundles, rendering the presence of linker molecules highly unlikely. Additionally, HRTEM images reveal that most PbSe QDs, which are attached to the SWNTs, grow with a preferred orientation of the {002} lattice planes perpendicular to the SWNT bundles. The influence of varying growth conditions was examined by changing the growth time and the Pb to Se ratio, thus identifying the optimal synthesis parameters to form PbSe-SWNT hybrids. These optimal parameters are a growth time of 5 min and a Pb to Se ratio of 1:1. Moreover it was found that the growth of PbSe QDs attached to SWNT bundles is favored over the growth of spherical particles in solution. Most probably this results from the stabilization of the PbSe QD dipole moment by the large electronic π-system of the SWNTs. The three-dimensional morphology of the hybrid, which was revealed by electron tomography, shows that the PbSe QDs use the SWNT bundles as a template for their growth. This results in QDs that exhibit the shape of half rings and thus differ from the classical spherical QD shape. Combining all experimental results, conclusions on the growth mechanism can be drawn: After an initial nucleation of the Pb-precursor on the SWNT sidewalls, the PbSe QD growth starts right after the Se-precursor injection. Further QD growth happens via spontaneous oriented attachment. The SWNT bundles are used as templates, this is why the PbSe QDs exhibit their characteristic half-ring shape. The application of PbSe-SWNT hybrids in photodetectors appeared to be very challenging, due to high dark currents generated by metallic SWNTs. However, the facile and promising synthesis initiated the use of other PbSe QD supporting materials, exhibiting non-metallic but at the same time high charge carrier transport

98

4.3 PbSe Quantum Dot-Layered Material Hybrids properties. Therefore, in the following section the previous established synthesis was transferred to a combination of PbSe QDs with layered materials, namely graphite/graphene and the TMDs MoS2 and WS2 .

4.3 PbSe Quantum Dot-Layered Material Hybrids 4.3.1 Fabrication of PbSe-Layered Material Hybrids Hybrids of PbSe QDs and layered materials were synthesized similar to PbSeSWNT hybrids, with an in situ hot-injection synthesis technique. A dispersion of either FLG, MoS2 , or WS2 was injected at 120 ◦ C into a stock solution of Pb-oleate, which was prepared of PbO, oleic acid, and 1-octadecene. The FLG, MoS2 , and WS2 dispersions were produced by several ultrasonication and centrifugation steps of the respective powder in diphenyl ether. Then, the temperature was raised to 170 ◦ C, followed by a Se-precursor injection to start the QD growth. After 5 min growth time the reaction was quenched. Pb2+ and Se2 – were used in great excess compared to the layered material in order to guarantee a high concentration of PbSe QDs and thus a dense coverage of the nanoflakes. Therefore, besides the QDs which grow on the nanoflakes, there are also many unbound PbSe QDs in solution. This is revealed by the brownish color of the dispersion. After several centrifugation steps with hexane, the PbSe-layered material was isolated. The nonpolar solvent hexane dissolves the PbSe QDs, whereas the hybrid precipitates in hexane, thus achieving a facile separation of both products (for details on synthesis see Experimental Part, Section 3.2). The BF-TEM images in Figure 4.15 were recorded after the separation process of the unbound QDs from the hybrids. It is obvious that in all three hybrid samples, namely PbSe-FLG, PbSe-MoS2 , and PbSe-WS2 , no unbound QDs are present after purification, clearly indicating the high efficiency of the separation process. EDX investigations were performed in order to verify the composition of each hybrid. Figure 4.16 shows that the spectra of all three hybrid structures exhibit the expected peaks, proving the presence of carbon, molybdenum/sulfur, or tungsten/sulfur besides PbSe. This shows that the synthesis approach, which was used to form PbSe-SWNT hybrids, also works for a wide range of layered materials. This promising result proves the general feasibility and the great potential of the formerly established synthesis. Moreover, the strong bond between QDs and supporting material, which was already found for PbSe-SWNT hybrids, was also observed for PbSe-layered material hybrids. Even a combination of ultrasonication and repeated purification cycles was not sufficient to separate the two materials. 99

4 Colloidal PbSe Quantum Dot Hybrid Materials (a) PbSe-FLG

(b) PbSe-MoS2

(c) PbSe-WS2

1 µm

200 nm

500 nm

(d) PbSe-MoS2

(e) PbSe-MoS2

20 nm

10 nm

Figure 4.15: BF-TEM images of (a) PbSe-FLG, (b) PbSe-MoS2 , and (c) PbSe-WS2 hybrids, showing a dense coverage of the nanoflakes with PbSe QDs. (d) and (e) HRTEM images of a PbSe-MoS2 hybrid, revealing PbSe and MoS2 crystal lattices and an average size of 5.7 nm of the PbSe QDs (corresponding to an absorption edge of 1675 nm and a band gap of 0.74 eV). (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

O Si Cu

1

Pb Se C

Fe Cu Fe

0 0

4

Cu

8

12

Energy (keV)

16

(c)

20

Pb Se Mo S

15 10 5 0

Si Fe Cu C O Cu Fe Cu

0

4

8 12 16 20

Energy (keV)

Counts (x103)

(b)

2

Counts (x103)

Counts (x103)

(a)

8 6

Si

Cu

C Fe

4

Cu

O

Pb Se W S

Cu

2

Fe

0 0

4

8

12

16

Energy (keV)

Figure 4.16: EDX spectra of (a) PbSe-FLG, (b) PbSe-MoS2 , and (c) PbSe-WS2 hybrids, showing the expected material peaks. Cu, Fe, Si, O, and (partly) C peaks can be ascribed to residuals from synthesis and elements arising from the measuring setup. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.3.2 Nanoflake Thicknesses of Layered Materials The shape of the layered material within the hybrid can easily be seen in the TEM images, however, the respective layer thicknesses are not visible. From the contrast in conventional BF-TEM images it can be deduced, that the layers are somewhat thinner than bulk material (Figure 4.17), but a more detailed investigation is necessary to come up with a precise statement.

100

4.3 PbSe Quantum Dot-Layered Material Hybrids (a) FLG

(b) MoS2

1 µm

500 nm

1 µm

10 nm-1

10 nm-1

10 nm-1

(c) WS2

Figure 4.17: BF-TEM images and corresponding SAED spectra of (a) FLG, (b) MoS2 , and (c) WS2 , indicating thin but no single layers. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

Different techniques like Raman spectroscopy, convergent beam electron diffraction (CBED), and cross-sectional HRTEM investigations were used to get the nanoflake thicknesses. The Raman spectrum of exfoliated graphite (with PbSe QDs on the surface) differs from the spectrum of untreated graphite flakes (Figure 4.18(a),(d)). PbSe itself is Raman inactive, and its oxidized form PbO exhibits strong Raman signals at 136 cm−1 and 274 cm−1 , and weaker ones at 176 cm−1 , 368 cm−1 , and 782 cm−1 . However, the displayed bands in all following spectra appear at different wavenumbers and therefore they can all be related to the exfoliated layered materials and not to the PbSe QDs.[531] Consequently, the differences in the two spectra in Figure 4.18(a) and 4.18(d) directly result from the exfoliation process. The exfoliated graphite shows a 2D band at 2690 cm−1 which appears to be relatively symmetric and higher in intensity than the G-band at 1580 cm−1 . Moreover, compared to the Raman spectrum of graphite, the 2D band of the exfoliated material is shifted by 30 cm−1 to smaller wavenumbers. Both of these facts clearly verify an exfoliation of graphite. However, the 2D band cannot be fitted with a single Lorentzian function with a full-width half maximum (FWHM) of 24 cm−1 , which indicates the presence of FLG rather than (monolayer) graphene (see Chapter 2, Section 2.3.1). Figure 4.18(b),(c),(e), and (f) display the Raman spectra of PbSe-MoS2 and PbSe-WS2 hybrids synthesized

101

4 Colloidal PbSe Quantum Dot Hybrid Materials with exfoliated MoS2 and WS2 and the spectra of MoS2 and WS2 powder. The corresponding spectra look alike, which means that both characteristic bands, namely E2g and A1g appear at the same wavenumbers for exfoliated material and powder (i.e., 383 cm−1 and ∼408 cm−1 for MoS2 and ∼351 cm−1 and 420 cm−1 for WS2 ). Differences in Raman spectra would appear between bulk TMDs and TMDs with stacks below 4 layers (see Chapter 2, Section 2.3.2).[328,347] Therefore it is assumed that all three layered materials are exfoliated to a certain degree, however no single-layers are achieved.

0.5 0.0 2000

0.4 0.0 2000

3000 -1

Raman shift (cm )

-1

383 cm-1

0.0 400

1.5

0.5

A1g = 409 cm E2g =

1.0 0.5

-1

383 cm-1

0.0 360

400

440 -1

Raman shift (cm )

420 cm-1 E2g = 351 cm-1

0.0

440

300

400

500

Raman shift (cm-1)

(f)

1.5

A1g =

1.0

Raman shift (cm-1)

Intensity (a.u.)

Intensity (a.u.)

0.8

2D = 2720 cm-1

408 cm E2g =

360

(e)

1.6 G= 1.2 1582 cm-1

0.5

(c)

A1g =

1.0

3000

Raman shift (cm-1)

(d)

1.5

Intensity (a.u.)

1.0

G= 1580 cm-1

(b)

Intensity (a.u.)

1.5

2D = 2690 cm-1

Intensity (a.u.)

Intensity (a.u.)

(a)

1.5 1.0

A1g = E2g = 420 cm-1 352 cm-1

0.5 0.0 300

400

500

Raman shift (cm-1)

Figure 4.18: Raman spectra of (a) PbSe-FLG, (b) PbSe-MoS2 , and (c) PbSe-WS2 hybrids and reference spectra of (d) graphite, (e) MoS2 powder, and (f) WS2 powder. A comparison between the exfoliated nanoflakes in (a),(b),(c) and the bulk materials in (e),(f),(g) indicate thin but no single layers in the hybrid structure. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

The thickness of exfoliated MoS2 nanoflakes was additionally investigated via CBED (Figure 4.19). For that, a drop of an exfoliated MoS2 dispersion was deposited on a lacey carbon copper TEM grid and identified by BF-TEM (Figure 4.19(a). In order to determine the flake thickness, an actual acquired CBED pattern and a simulated CBED pattern obtained for identical conditions are compared in order to identify the best match and thus draw a conclusion on the thickness of the MoS2 nanoflake. The CBED pattern was acquired in a two-beam condition with the direct (0000) beam and the Bragg-diffracted (-1-120) beam. The best match of several acquired and simulated CBED patterns was found for MoS2 flakes of a few tens of nm (the example of Figure 4.19 resulted in 41 nm). Therefore, in terms of electrical properties, the nanoflakes can be considered as bulk semiconductors,

102

4.3 PbSe Quantum Dot-Layered Material Hybrids however they are still solution-processable.[532] Moreover, CBED investigations reveal a high crystallinity of the MoS2 nanoflakes, indicating that the exfoliation process does not significantly damage the MoS2 crystal lattice. (0000) (b)

(-1-120)

(c) 200 nm

experiment simulation

(a)

Figure 4.19: (a) BF-TEM image and corresponding (b) CBED pattern of an exfoliated MoS2 nanoflake. (c) Simulated CBED pattern for a MoS2 nanoflake of 41 nm, showing a good match with the experimentally obtained CBED pattern in (b). (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.3.3 PbSe-MoS2 Interface Characterization Similar to the PbSe-SWNT hybrids, the interface between PbSe and the layered material is the most important and critical part regarding a possible application of the hybrids in photodetectors. Electron tomography investigations as for PbSeSWNT hybrids are not feasible for PbSe-layered material hybrids. The layered material would increase in projected thickness, while tilting the TEM grid around a single axis. This would impede decent imaging conditions and thus render a three-dimensional reconstruction of the PbSe QDs inaccurate. Therefore, SAED analysis in combination with top-view and cross-sectional HRTEM investigations were used to take a closer look at the interface and to draw conclusions on a possible growth mechanism. All of the following studies were exemplary made for one PbSelayered material hybrid, namely PbSe-MoS2 . This hybrid exhibits a very promising structure, because its material combination belongs to the so-called misfit-layered compound structure (see Chapter 2, Section 2.4.3). For the same crystal structure combination as the investigated hybrid material, namely thin-films of PbS grown by molecular beam epitaxy on freshly cleaved TiS2 (0001) substrates, Spiecker et al. were already able to observe the misfit-layered compound structure.[533] This particular structure makes an epitaxial growth and thus a direct interface between PbSe QDs and MoS2 nanoflakes highly probable. Figure 4.20(a) shows a SAED pattern of randomly oriented PbSe QDs drop-cast from solution onto a TEM grid and 4.20(b) depicts a SAED pattern of exfoliated

103

4 Colloidal PbSe Quantum Dot Hybrid Materials MoS2 nanoflakes (without PbSe QDs). The SAED pattern of the PbSe QDs exhibits the expected rock-salt reflections, which appear in continuous rings with constant intensity rather than in distinct point reflections, due to a statistically distributed orientation of the PbSe QDs. The MoS2 reflections in Figure 4.20(b) emerge in a hexagonal arrangement for an electron illumination perpendicular to the basal plane (i.e., along h0001iMoS2 ). If PbSe QDs were randomly oriented on the MoS2 nanoflakes, the SAED pattern of the PbSe-MoS2 hybrid would be a simple combination of both single spectra, however, the measured diffraction pattern, which is depicted in Figure 4.20(c), differs from that simple combination. The hexagonal pattern caused by the MoS2 crystal structure is the same as in Figure 4.20(b), but the PbSe diffraction pattern differs from the PbSe-only pattern. The reflection rings are no longer continuous rings, they are cut in twelve equally spaced ring fragments on the respective {200}P bSe and {220}P bSe diffraction ring positions. The ring fragments of the inner ring correspond to the {200}P bSe reflections, while the ring fragments of the outer ring are generated by {220}P bSe reflections. Figure 4.20(c) shows that the inner and the outer ring fragments are rotated by 45◦ with respect to each other. Moreover, the pattern depicts no diffraction rings of type {111}P bSe and {311}P bSe . Both of these findings lead to the conclusion that the PbSe QDs attached to the MoS2 nanoflakes grow in the h001iP bSe direction (i.e., PbSe(001)kMoS2 (0001)) and exhibit a preferred in-plane orientation (i.e., PbSe(200)kMoS2 {1-100}). These two observations are clear evidences for an epitaxial relationship between the PbSe QDs and the MoS2 nanoflakes. (a) PbSe

(b) MoS2

(c) PbSe-MoS2 (220)PbSe _ (1120)MoS2

_ (1100)MoS2

(200)PbSe

(020)PbSe

5 nm-1

5 nm-1

5 nm-1

Figure 4.20: SAED pattern of (a) PbSe QDs, (b) a MoS2 nanoflake, and (c) in situ synthesized PbSe-MoS2 hybrids. In (c) PbSe QD reflections appear in distinct ring fragments rather than continuous rings, revealing a preferred orientation of PbSe QDs on the MoS2 nanoflakes represented by the red, green, and yellow squares. The white hexagon indicates the hexagonal arrangement of MoS2 . (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

One possible explanation for this preferred growth direction and the associated strong bonding of PbSe QDs on the MoS2 nanoflakes might be the presence of

104

4.3 PbSe Quantum Dot-Layered Material Hybrids dipole-dipole interactions between the two components. Although PbSe QDs exhibit a centrosymmetric crystal structure, around 89 % of all PbSe QDs have a dipole moment. Statistically, the largest dipole is along the h001iP bSe axis (see Chapter 2, Section 2.1.4 and Cho et al.[61] ). Figure 4.21 schematically depicts the prevailing arrangement of the PbSe QDs with the different surface facets on the first layer of a MoS2 nanoflake. The red arrows are marked perpendicular to the surface of the MoS2 nanoflake and indicate the h100iP bSe axis and at the same time the direction of the dipole moment. The green (facets terminated by Pb atoms) and gray (facets terminated by Se atoms) triangles represent the polar {111}P bSe facets, while the remaining PbSe facets are colored in white and indicate the polar {100}P bSe facets. PbSe

MoS2 S

S Mo

S

S Mo

S

S Mo

S

S Mo

S

S Mo

S

S Mo

S

S Mo

S

S

Figure 4.21: Schematic illustration demonstrating the perpendicular orientation of the PbSe QD dipole moment (red arrow) to the MoS2 nanoflake. Green and gray areas: polar {111} facets terminated either by Pb or by Se atoms; white areas: nonpolar {100} facets. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

The schematic illustration in Figure 4.22(a) shows the arrangement of Pb, Se, and S atoms in real space. Due to symmetry considerations there are three equivalent orientation variants of PbSe QDs on the MoS2 nanoflakes, which are rotated by 120◦ with respect to each other. In Figure 4.20(c) each of the three orientation variants is represented by a square (green, yellow, or red) interconnecting the reflections of one variant (reciprocal space). The formerly assumed bulk misfitlayered compound structure of PbSe QDs (monochalcogenide) and MoS2 nanoflakes (dichalcogenide) is finally proven by the observed orientation relationship including the three orientation variants.[534] However, in contrast to perfect misfit-layered compound structures like the one made of PbS and TiS2 ,[533] where both lattices perfectly match, there is a huge lattice misfit of 12 % between the adjacent bSe 2 lattice planes of PbSe (dP200 = 0.3064 nm [ICSD no. 38294]) and MoS2 (dMoS 1−100 = 0.2737 nm [ICSD no. 84180]). The lattice misfit can be seen in the PbSe-MoS2 SAED pattern, where PbSe and MoS2 reflections are spatially separated and do

105

4 Colloidal PbSe Quantum Dot Hybrid Materials not superimpose (Figure 4.20(c)). Due to this huge lattice misfit it is assumed that this direction is not the main reason for the epitaxial growth of the PbSe QDs on the MoS2 nanoflakes. However, in the perpendicular in-plane direction, i.e., along h11-20iMoS2 , the PbSe (d400 = 0.1532 nm) and MoS2 (d11−20 = 0.1580 nm) lattice planes show a reasonable good fit with a lattice mismatch of only 3 %, allowing an epitaxial growth, which results in the observed preferred orientation relationship. As depicted in the illustration in Figure 4.22(b), a lattice mismatch of 3 % is still large enough to leave the PbSe QDs a certain remaining rotational degree of freedom on the MoS2 nanoflakes, generating the observed ring fragment reflections in the SAED pattern rather than discrete diffraction spots. The rotational range was directly determined from the SAED pattern and does not exceed 12◦ for most of the PbSe QDs.

Figure 4.22: (a) Schematic illustration of the arrangement of Pb (“gray”), Se (“green”), and S (“red”) atoms, showing that the three equivalent orientation variants of PbSe QDs on the MoS2 nanoflakes are rotated by 120◦ with respect to each other. (b) Schematic illustration showing that the rotational degree of freedom of the PbSe QDs on the MoS2 nanoflakes (real space) results in an elongation of the diffraction spots (reciprocal space). (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

The epitaxial growth is not only visible in the SAED patterns of PbSe-MoS2 hybrids, but also in the corresponding in-plane HRTEM images. The contrast in the HRTEM image in Figure 4.23(a) is dominated by the hexagonal lattice of a MoS2 nanoflake, because PbSe QDs are small compared to the large MoS2 nanoflake. Nevertheless, PbSe QDs are clearly visible and every QD can be assigned to one of the three orientation variants (encircled in red, green, or yellow), clearly indicating the preferred growth orientation of the QDs. Figure 4.23(b) depicts enlarged HRTEM images of representative PbSe QDs of each orientation variant. The HRTEM contrast shows a stripe-like modulation along one of the three h1-100iMoS2 lattice directions, wherever the MoS2 flake is covered with PbSe QDs. This modulation is generated by an interference of electrons, which are scattered in the PbSe and MoS2 lattices. Due to a multiple scattering resulting

106

4.3 PbSe Quantum Dot-Layered Material Hybrids from various spatial frequencies along the mismatched h1-100iMoS2 direction, the modulation contrast is more complex than the modulation expected from a simple Moir´e interference. (a)

(b) 1

3 2 10 nm

1

2

3

Figure 4.23: (a) HRTEM image of PbSe-MoS2 hybrids, showing that every QD can be assigned to one of the three orientation variants (encircled in red, green, or yellow). (b) Magnified HRTEM images of representative PbSe QDs of each orientation variant (position 1, 2, and 3 in (a)). (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.3.4 Control Experiments with Ex Situ Fabricated Hybrids In order to find out if an in situ synthesis is necessary to achieve a preferred orientation of PbSe QDs on MoS2 nanoflakes or if a simple mixing of both components is enough to induce a ligand exchange between oleic acid and MoS2 , ending up with the same result as the in situ synthesis, two control experiments were performed. In a first approach pre-synthesized PbSe QDs were mixed with exfoliated MoS2 nanoflakes by a simple stirring for 1 hour at room temperature. In a second approach both components were mixed and heated to 170 ◦ C under argon atmosphere and constant stirring. After four hours the product was isolated and investigated by BF-TEM and SAED. Figure 4.24 shows that both experiments lead to an assembly of PbSe QDs on MoS2 nanoflakes. However, SAED investigations clearly reveal the absence of a preferred orientation of the QDs on the MoS2 nanoflakes. The SAED patterns in Figure 4.24 show that the reflections resulting from the PbSe lattice form complete rings rather than distinct diffraction spots or ring fragments, indicating a random distribution of the PbSe QDs on MoS2 . These results clarify that first, only an in situ synthesis is able to lead to a preferred orientation of PbSe QDs on MoS2 nanoflakes and second, during the in situ synthesis PbSe QDs directly grow on the MoS2 nanoflake rather than being attached after formation.

107

4 Colloidal PbSe Quantum Dot Hybrid Materials (a)

50 nm

5 nm-1

(b)

100 nm

5 nm-1

Figure 4.24: BF-TEM images and corresponding SAED pattern of PbSe-MoS2 hybrids prepared by mixing pre-synthesized PbSe QDs with exfoliated MoS2 nanoflakes (a) under constant stirring for 1 h at room temperature and (b) upon heating to 170◦ C for 4 h under constant stirring and an argon atmosphere. Although BF-TEM images reveal a hybrid structure, SAED pattern show no preferred orientation of PbSe QDs. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.3.5 PbSe-MoS2 Interface Visualization So far all evidences for a direct contact between PbSe QDs and MoS2 nanoflakes are indirect proofs, which leave room for speculations. A direct visualization of the interface between PbSe QDs and MoS2 nanoflakes would strengthen the assumptions and provide an unambiguous proof for the absence of linker molecules. Therefore, a cross-sectional TEM sample of a MoS2 nanoflake covered with PbSe QDs was prepared by focused ion beam (FIB) and characterized by BF-TEM and HRTEM. The TEM images in Figure 4.25 show the cross-section of this PbSe-MoS2 hybrid. The contrast clearly indicates the layered structure of the MoS2 flake, with 2 an interlayer distance of dMoS 0002 = 0.6145 nm [ICSD no. 84180]. Moreover it can be seen that the layered structure of the MoS2 flake is slightly damaged, which might be caused by milling and electron beam irradiation in the course of cross-sectional sample preparation. The representative MoS2 nanoflake exhibits a thickness of around 25 nm, which is in good agreement with former CBED characterization results for layer thickness (see Figure 4.19). The cross-sectional image clearly shows that the whole MoS2 surface is densely covered with PbSe QDs. The QDs

108

4.3 PbSe Quantum Dot-Layered Material Hybrids are intimately attached to both sides of the nanoflake. In the HRTEM image in Figure 4.25(b) it is obvious that there is no space between the two materials, this is why the presence of oleic acid (length of an oleic acid molecule is around 2 nm) or any other linker molecule or impurity at the interface can be excluded. This intimate contact between PbSe QDs and MoS2 nanoflakes was already expected due to the identification of a preferred orientation of the QDs on the nanoflakes. However, cross-sectional BF-TEM and HRTEM investigations finally provide an unambiguous proof of an epitaxial growth resulting in a direct and ligand-free interface. Cross-sectional TEM images of the PbSe-MoS2 hybrids are not only extremely helpful to proof the intimate contact between PbSe and MoS2 , they also confirm the preferred crystallographic orientation of the PbSe QDs on the MoS2 nanoflakes, which was already observed in SAED patterns and HRTEM images in plan-view geometry. The two representative QDs in Figure 4.25(c) show that the {200}P bSe lattice planes are aligned with the {11-20}MoS2 lattice planes of the MoS2 substrate, indicating a preferred growth of the QDs along the [001]P bSe axis. (a)

(c)

PbSe

Carbon

PbSe d200

[001]PbSe

MoS2 30 nm

Si

(b) PbSe MoS

5 nm

MoS2

MoS2 d0002

2 nm

d11-202

Figure 4.25: (a) BF-TEM image of a cross-sectional PbSe-MoS2 hybrid on a silicon substrate, covered with a protective carbon layer. The BF-TEM image shows that the MoS2 nanoflake is densely covered with PbSe QDs on both sides. (b) HRTEM image revealing an intimate contact between PbSe QDs and a MoS2 nanoflake (labeled in orange). (c) Enlarged HRTEM image demonstrating the preferred crystallographic orientation of the PbSe QDs along the [001]P bSe axis (blue arrow), which is reflected by the nearly parallel alignment of {200}P bSe and {11−20}MoS2 lattice planes. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.3.6 Ligand Arrangement The observation of a direct contact between PbSe QDs and MoS2 nanoflakes implies that there are no linker molecules connecting PbSe QDs and MoS2 nanoflakes. However, any information about the outer surface of the QDs, which is not in contact with the MoS2 nanoflake, is still missing. Usually, semiconductor QDs

109

4 Colloidal PbSe Quantum Dot Hybrid Materials synthesized via hot-injection are covered with ligands in order to be stabilized in solution. A characterization of the PbSe-MoS2 hybrid by FT IR spectroscopy shows aliphatic stretching modes at 2850 cm−1 and 2916 cm−1 , indicating the presence of alkyl groups (Figure 4.26). Since no post-synthesis ligand exchange was performed, these alkyl groups have to be assigned to a molecule, which was used during the synthesis process, reducing the options to oleic acid and TOP. It is well known from literature that first, oleic acid preferentially binds to a Pb-terminated surface, while TOP binds to a Se surface and second, PbSe QDs are usually Pb-terminated.[192,198] These facts render the presence of oleic acid ligands more probable than the presence of TOP ligands. Moreover, for PbSeSWNT hybrids synthesized under the same conditions than PbSe-MoS2 hybrids (see Section 4.2.2), the presence of phosphorous was excluded by means of XPS. Additionally, HRTEM and SAED exclude any molecule, including oleic acid, at the interface. Consequently, it is assumed that oleic acid is only present at the outer surface of the PbSe QDs and not at the interface between PbSe QDs and MoS2 nanoflakes. This type of ligand arrangement was formerly reported by Huang et al.[396] and also found for PbSe-SWNT hybrids (see Section 4.2). Thus, on the outside, oleic acid stabilizes the hybrid and protects the QDs against surface oxidation and other environmental influences, while at the ligand-free and direct interface between PbSe QDs and MoS2 nanoflakes charge transport can happen on a fast timescale without being affected through long and insulating organic molecules. Therefore, PbSe-MoS2 hybrids need no time consuming and laborious ligand exchange to guarantee a fast charge transfer. A ligand removal or exchange often has the disadvantage of creating trap states. Trap states are possible charge recombination sites, which can withdraw charge carriers that could otherwise contribute to a photocurrent generation. Therefore, PbSe-MoS2 hybrids appear to be very promising for electronic applications.

4.3.7 Growth Mechanism Combining all the experimental results and taking into account previous reports on the growth properties of pure PbSe QDs,[59,61,195,528] PbSe QDs on SWNTs (see Section 4.2 and Chen et al.[529] ), and the chemical structure of misfit-layered compounds,[398,400] it is possible to propose a growth mechanism for the epitaxial growth of PbSe QDs on MoS2 nanoflakes. This mechanism is similar to the growth mechanism for PbSe-SWNT hybrids (introduced in Section 4.2.5). Figure 4.27 shows, that the concentration of PbSe QDs on TMD nanoflakes is

110

Transmission (%)

4.3 PbSe Quantum Dot-Layered Material Hybrids 100

98

2850 cm-1

96 3200

2916 cm-1 3000

2800

2600 -1

Wavenumber (cm ) Figure 4.26: FT IR spectrum of PbSe-MoS2 hybrids, showing aliphatic stretching modes at 2850 cm−1 and 2916 cm−1 , indicating alkyl groups in the sample. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

always higher than on FLG. This observation can be explained as follows: QD growth starts with a coordination of lead oleate directly on the surface of the nanoflakes. Single layers of TMDs (e.g., MoS2 , WS2 ) are composed of a sandwich structure with a S-X-S (X = Mo, W) stoichiometry, providing a negative surface polarization due to sulfide termination (see Chapter 2, Section 2.3.2 for details). Thus, Coulomb interactions between Pb2+ cations and S2 – anions initiate the coordination of lead oleate onto the TMD nanoflakes. In comparison, FLG only provides a large conjugated π-electronic system to compensate the charges of the Pb2+ cations. Therefore, more lead precursors are attracted to TMDs than to FLG. The nucleation step is followed by the injection of the selenium precursor, initiating the PbSe QD growth on the nanoflakes as well as in solution. It is assumed that the amount of lead precursor attracted by the nanoflake influences the number of PbSe QDs grown on a nanoflake. Since the density of PbSe QDs on TMDs is higher than on FLG, it can be deduced that an effective coordination and nucleation of lead on the nanoflake is crucial to start the QD growth on the nanoflake, and thus to end up with a high concentration of PbSe QDs on the nanoflakes. The epitaxial growth of PbSe QDs on MoS2 nanoflakes (supported by BF-TEM, HRTEM, and SAED investigations) leads to a very strong bonding between the QDs and the nanoflakes, which cannot be broken even by repeated cycles of ultrasonication and centrifugation. Usually, the surface of PbSe QDs is covered with ligands, which prevent any charge transport from and to the QDs and therefore would make the hybrid useless in terms of photodetector applications.[119,392] However, the realized epitaxial growth implies a direct and linker-free interface, which enables a fast charge transfer from the PbSe QDs to the nanoflakes and thus is crucial for the application of PbSe-nanoflake hybrids in photodetectors.

111

4 Colloidal PbSe Quantum Dot Hybrid Materials (a)

20 nm

PbSe-MoS2

250 nm

(b)

50 nm

250 nm

PbSe-WS2

250 nm

(c)

50 nm

250 nm

250 nm

250 nm

PbSe-FLG

250 nm

500 nm

250 nm

Figure 4.27: Representative HRTEM and BF-TEM images of (a) PbSe-MoS2 , (b) PbSe-WS2 , and (c) PbSe-FLG, revealing a denser QD coverage of TMD flakes than of FLG. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.3.8 Summary PbSe-layered material hybrids were successfully produced via an in situ hotinjection synthesis. The general feasibility of the synthesis was proven by fabricating three different material combinations, namely PbSe-FLG, PbSe-WS2 , and PbSe-MoS2 . The whole fabrication process follows a wet-chemical synthesis route, this is why the hybrid material remains solution-processable throughout the whole process and thus is suitable for a wide variety of applications. All three hybrids were characterized by EDX, BF-TEM, and HRTEM, showing the chemical compo-

112

4.4 PbSe-MoS2 Photodetectors sition and the arrangement of PbSe QDs on the nanoflakes. Raman spectroscopy further indicated an exfoliation of graphite, bulk MoS2 , and bulk WS2 to few-layer nanoflakes. Subsequently, a more detailed and extensive characterization was exemplarily performed for PbSe-MoS2 hybrids. For these hybrids an epitaxial growth is highly probable, since PbSe-TMDs belong to the class of misfit-layered compounds. The expected epitaxial growth was experimentally investigated via SAED, top-view, and cross-sectional HRTEM characterization. The results gained from SAED characterization and HRTEM investigations corroborate a preferred orientation of PbSe QDs on MoS2 nanoflakes, with three equivalent orientation variants, rotated by 120◦ relative to each other. Moreover, the direct and linker-free interface between PbSe QDs and MoS2 nanoflakes was clearly shown by cross-sectional HRTEM investigations. However, FT IR spectroscopy identified the presence of molecules with alkyl chains in the hybrid sample. On the basis of all results it is assumed that oleic acid ligands only protect the outer surface of the PbSe QDs, while the interface between PbSe QDs and MoS2 remains linker free. Therefore, PbSe QDs are well protected against oxidation and other surface modifications, while the direct interface between PbSe and MoS2 is ideal for a fast charge transfer. In conclusion, the air-stable PbSe-MoS2 hybrids need no laborious and trap-state inducing ligand-exchange steps after synthesis in order to be used for solutionprocessing. In fact the material is immediately and without further modification ready-to-use for application in NIR photodetectors.

4.4 PbSe-MoS2 Photodetectors 4.4.1 Basic Photodetector Characterization Photodetectors were fabricated by drop-casting a hexane dispersion of PbSeMoS2 on glass substrates with pre-patterned interdigitated gold electrodes (Figure 4.28). All devices were investigated under ambient air conditions, without being encapsulated. For photoresponse characterization a broadband halogen light source with a 1200 nm cut-off filter and an incident power of about 1.6 mW was used. The average diameter of PbSe QDs on MoS2 nanoflake is 5.7 nm (see HRTEM image in Figure 4.15(d)). This results in a band gap of 0.74 eV, corresponding to an absorption edge of 1675 nm.[197] The PbSe QDs are responsible for the photosensitization of the MoS2 nanoflakes, while the MoS2 nanoflakes are used as charge transporting layer. The charge carrier mobility in MoS2 nanoflakes is

113

4 Colloidal PbSe Quantum Dot Hybrid Materials

100 µm

50 µm

1 µm

100 nm

Figure 4.28: SEM images of PbSe-MoS2 hybrids on pre-patterned interdigitated gold electrodes (channel width = 5 µm). The last image reveals the PbSe QDs on the MoS2 nanoflake. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

much higher than in PbS QD thin-films, even if the flakes are overlapping each other. As introduced in the previous section, MoS2 nanoflakes exhibit a few-layer rather than single-layer structure. Consequently, in terms of optical properties, MoS2 can be considered as a bulk material. Therefore, the band gap of the MoS2 nanoflakes in the hybrid structure is 1.2 eV (corresponding to 1030 nm), this is why MoS2 would absorb light with photon energies above 1.2 eV. In order to assure that MoS2 nanoflakes are only responsible for charge transport, a 1200 nm cut-off filter was used to avoid any photoresponse from the MoS2 nanoflakes. Upon illumination, PbSe QDs are excited and an exciton is formed. In order to achieve exciton dissociation at the PbSe-MoS2 interface, a bias was applied to the electrodes, resulting in a separation of the electron-hole pair. Usually, exciton dissociation is characterized via time-dependent photoluminescence (PL) quenching measurements.[384,535] However, PL from the PbSe QDs is beyond the detector limit (InGaAs ≤ 1600 nm) and thus cannot be detected. Consequently, exciton separation can only be indirectly supported by photocurrent generation. Figure 4.29 depicts the photoresponse of two different photodetectors. The first photodetector was made of pristine MoS2 nanoflakes, while the second photodetector was fabricated with PbSe-MoS2 hybrids. The photoresponse was measured under a constant applied bias as a function of illumination (λ > 1200 nm). The PbSe-MoS2 photodetector shows a clear photo-switching behavior in response to the incident illumination, while the MoS2 photodetector exhibits no photoresponse, and thus demonstrates that light absorption only results from the PbSe QDs. Most probably the photocurrent is generated by a transfer of electrons from the PbSe QDs to the MoS2 nanoflakes. This assumption is based on the observation of a current increase rather than a decrease upon illumination. After an electron transfer, the charge carrier density in the initially n-type MoS2 nanoflakes[318,536] is increased leading to a resistance decrease and therefore a photocurrent increase. The overall responsivity in the NIR region was calculated to be 1.9 µA/W. However, the incident optical power could not be determined accurately and is probably overestimated, which would result in an underestimated responsivity. Therefore, the value of 1.9 µA/W only represents a lower boundary of the photoresponse. 114

4.4 PbSe-MoS2 Photodetectors

Current (nA)

4

ON

OFF

ON

OFF

ON

OFF

ON

3 PbSeMoS2

2 1

MoS2

0

430 ms

250 ms

0

20

40

60

80

Time (sec) Figure 4.29: Photoresponse of a PbSe-MoS2 photodetector (black line) upon NIR illumination (λ > 1200 nm), with response times of 250 ms into the on-state and 430 ms into the off-state. The red line depicts the photocurrent of a MoS2 photodetector upon illumination with the same light source, showing no photoresponse. The ON-OFF switching time of the light source was 10 s, Vbias = 14 V. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.4.2 Repeated Switching Stability

Current (nA)

Regarding a future application of the hybrid in photodetectors, not only the photoresponse but also the stability of the device upon many switching cycles is important. In Figure 4.30 it can be seen that the photoresponse shows no considerable changes in signal strength and shape even after 48 cycles of switching the illumination source on and off. Therefore, it can be stated that the investigated PbSe-MoS2 photodetector is stable during operation in air and under illumination. 2 1 0 200

300

400

500

600

Time (sec) Figure 4.30: Photoresponse of a PbSe-MoS2 photodetector, demonstrating the photocurrent stability over 48 switching cycles. The photodetector exhibits an excellent photocurrent stability. The ON-OFF switching time of the light source was 5 s, Vbias = 10 V. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.4.3 Dark Current An important figure of merit of a photodetector is its dark current. The dark current of PbSe-MoS2 detectors was measured without illumination under a constant applied bias. Figure 4.31(a),(b) shows that the dark current initially increases

115

4 Colloidal PbSe Quantum Dot Hybrid Materials over time until it saturates at a relatively high value in the range of hundreds of µA. This behavior can only be observed for PbSe-MoS2 photodetectors, while the dark current of MoS2 photodetectors is substantially lower and does not increase over time (see Figure 4.31(c)). It is possible to explain this observation by taking a closer look at the structural composition of misfit-layered compounds. Brandt et al. explained, that there is a charge transfer from the MX part to the TX2 layer.[401,402] In the PbSe-MoS2 hybrid, PbSe represents the monochalcogenide MX, while MoS2 corresponds to the dichalcogenide TX2 . With this in mind, the PbSe QDs probably dope the MoS2 nanoflakes even before illumination. However, the doping reaches a steady-state at a certain value, resulting in a stabilization of the dark current. After stabilization, the increase and decrease characteristics of the photocurrent are symmetric. It is important to note that in all photoresponse spectra of this Chapter, the dark current was subtracted in order to get a constant baseline.

0.26 light ON light OFF

0.24

0.22

0

100 200 300

(c)

0.14 0.13

light ON light OFF

0.12 0.11

Time (sec)

0

200 400 600

Time (sec)

Current (nA)

(b) Current (µA)

Current (µA)

(a)

30

20 light OFF & ON

10 0

200

400

Time (sec)

Figure 4.31: Dark currents and photocurrents of a PbSe-MoS2 photodetector under illumination (λ > 1200 nm) with a bias of (a) 14 V and (b) 10 V and (c) of a MoS2 photodetector with a bias of 14 V (“black”) and 10 V (“blue”). The dark current of the hybrid photodetector increases with time until a high saturation current (µA), while the dark current of the MoS2 photodetector is much lower (nA) and shows no increase over time. ((a) and (b) adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.4.4 Photoresponse Time As marked in Figure 4.29, the increase as well as the decrease of the PbSe-MoS2 photocurrent consist of two-components: an initial, fast rising component from 0 % to 75 % of the maximum signal, and a second, much slower component from 75 % to 100 %. The fast component indicates the photoresponse time of the detector. A switch-on time of 160-250 ms and a switch-off time of 430-600 ms were extracted from Figure 4.29. The slow component might result from defects or trap states at the surface of the QDs.[352,461] The photoresponse time of PbSeMoS2 photodetectors is most probably determined by the charge carrier mobility

116

4.4 PbSe-MoS2 Photodetectors within the MoS2 nanoflakes rather than by the charge carrier separation at the interface, because the exciton dissociation happens on an ultrafast timescale. The photoresponse times of the here presented PbSe-MoS2 hybrid photodetectors are slower than of PbS-graphene photodetectors, which were reported by Konstantatos et al.[352] However, charge carrier mobility in MoS2 flakes is much slower than in graphene (see Chapter 2, Section 2.3), this is why PbSe-MoS2 hybrid photodetectors should be compared to MoS2 photodetectors, rather than to photodetectors based on graphene. Yin et al. measured the photoresponse time of a large MoS2 monolayer to be 50 ms,[537] showing that the photoresponse time of the here presented PbSe-MoS2 detector is slower but still in the same order of magnitude. However, in comparison to MoS2 photodetectors, PbSe-MoS2 hybrid detectors provide the big advantage of a wavelength-tunable detection range. Moreover, in this report solution-phase exfoliated MoS2 was used compared to mechanically exfoliated MoS2 by Yin et al.[537] and mechanically exfoliated graphene used by Konstantatos et al.[352] Solution-phase exfoliation results in smaller flakes than mechanical exfoliation, therefore more hopping events are necessary to be able to extract electrons at an electrode, possibly explaining the slower photoresponse. It has to be noted that the halogen light source was manually turned on and off in order to switch the illumination instead of using a shutter. The halogen light source might exhibit fluctuations right after turning it on, which was not taken into account during data evaluation. Therefore, extracted photoresponse times should be seen as an orientation to roughly classify the device, but not as exact values.

4.4.5 Applied Bias Dependence In order to investigate the dependence of the photoresponse on the applied bias, Vbias was constantly increased from 4 V to 18 V in steps of 2 V. Figure 4.32(a) indicates that the photocurrent increases with increasing bias. This observation is explained by a suppressed charge recombination and enhanced exciton dissociation n with increasing bias. The double logarithmic plot in Figure 4.32(c) shows a Vbias relation of the measured photocurrent with applied bias, where n is approximately 1.5. This finding was unexpected, because it differs from results of a similar structure. Konstantatos et al. measured a linear dependence of the photocurrent on the applied bias for PbS-graphene hybrids.[352] The photocurrent is calculated by subtracting the dark current from the total current under illumination. Thus, 1.5 the measured increase of Vbias in the PbSe-MoS2 hybrids might be explained by a

117

4 Colloidal PbSe Quantum Dot Hybrid Materials relatively stronger increase of the total current compared to the dark current. (a) 18 V

Current (nA)

6

 2V

4 2

4V

0 -4

0

4

8

12

Time (sec)

(c) log(Current)

Current (nA)

(b) 6 4 2 0 4

8 12 16 20

Bias (V)

1

5

10 1520

log(Bias)

Figure 4.32: (a) Photoresponse of a PbSe-MoS2 photodetector at different bias voltages, varied from 4 V to 18 V in steps of 2 V. (b) Current-bias characteristics and the corresponding (c) log(∆Current)-log(Bias) plot with linear fit (black line), which indicates a Vnbias relation of signal with bias with n ∼ 1.5 for voltages > 4 V. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

4.4.6 Photoresponse Air-stability All results reported so far demonstrate that the devices are air-stable without any encapsulation. Even after weeks and numerous illumination and switching cycles, the PbSe-MoS2 photodetectors showed almost no degradation. This remarkable stability can be explained by the already identified ligand arrangement. The outer surface of the highly air-sensitive PbSe QDs is passivated by oleic acid molecules, which exhibit a long and insulating aliphatic chain, whereas the interface is free of any ligands or linker molecules due to the direct and epitaxial growth of PbSe QDs on the MoS2 nanoflakes. Oleic acid protects the PbSe QDs from oxidation and prevents the formation of charge quenching sites, while charge separation can still occur at a fast timescale at the direct interface. Therefore, the big advantage of using epitaxial grown PbSe-MoS2 hybrids in photodetectors, compared to using only PbSe QDs or non-epitaxially grown hybrids, is the avoidance of a postsynthesis ligand removal or exchange. Usually, this post-synthesis step is necessary

118

4.4 PbSe-MoS2 Photodetectors to achieve charge transport, but it is time-consuming and often creates surface trap states.[352,386]

4.4.7 Control Devices with Ex Situ Fabricated Hybrids In order to corroborate the previous statements, two control devices were fabricated: For the first device as-synthesized PbSe QDs were drop-cast on a pre-patterned glass substrate without any further treatment, while for the second device presynthesized PbSe QDs were mixed with exfoliated MoS2 nanoflakes by stirring at room temperature for 1 h. As already expected, the PbSe-only device exhibited no photoresponse. PbSe QDs are covered with insulating oleic acid ligands, preventing any charge transport. The second control device showed some photoresponse, however it was weaker (photoresponse of 0.1-0.2 nA at a Vbias of 10 V) than the photoresponse of the in situ synthesized PbSe-MoS2 hybrids. In Section 4.3.4 it was shown that a simple mixing of pre-synthesized PbSe QDs with exfoliated MoS2 nanosheets does not result in a preferred orientation of the QDs on the nanoflakes, thus concluding that with a high probability there are oleic acid capping ligands between PbSe QDs and MoS2 nanoflakes. These ligands lead to an inhibited charge separation and an inferior photoresponse.

4.4.8 Flexible Photodetectors PbSe-MoS2 hybrids are solution-processable and therefore it was possible to fabricate flexible photodetectors. These photodetectors were prepared identically to rigid photodetectors, with the only difference of drop-casting the PbSe-MoS2 hybrid onto a flexible pre-patterned PET substrate rather than onto a glass substrate. Flexible photodetectors were characterized in three different positions, namely unbent, bent inward, and bent outward. Thereby, attention was paid on bending the device always perpendicular to the electrodes. The characterization was carried out under a constant applied bias of 10 V and apart from that under the same conditions than rigid substrates, using the identical broadband halogen light source and the 1200 nm cut-off filter. First, the device was bent inward, bringing the active layer (i.e., PbSe-MoS2 ) under compressive stress. Secondly, the photodetector was bent outward resulting in the active layer under tensile stress. Figure 4.33 displays the photocurrents for both situations, showing a photocurrent increase during inward bending (Figure 4.33(a)), while a photocurrent decrease was measured when the device was bent outward (Figure 4.33(b)). However, every time the flexible device was bent back to its flat position the photocurrent recovered to

119

4 Colloidal PbSe Quantum Dot Hybrid Materials the initial photocurrent, verifying that both, inward as well as outward bending are reversible processes. It is remarkable that the flexible PbSe-MoS2 photodetector shows a clear and stable photoresponse up to the smallest measurable bending radii of r = 3.8 mm for inward bending and r = 2.9 mm for outward bending. Both maximum bending radii are limited by the characterization setup and not by a failure of the device. In order to use PbSe-MoS2 hybrids in flexible photodetectors, not only the maximum bending radii but also the number of bending cycles is crucial. Therefore, the device performance was measured in a flat position after repeated bending. Figure 4.33(c) displays the photoresponse after bending the device 100, 200, 300, and 400 times. At first, the photodetector was bent outward for 200 times to a bending radius of 4.4 mm, then it was bent inward for another 200 times to a bending radius of 4.3 mm. The photocurrent only shows a small decrease, but the overall photoresponse characteristics remain stable during all bending cycles. (a)

(b)

(c) 1.2

0.8

0.0 0

200

400

Time (sec)

r = 3.8 mm

3

r 0.6

0.0 0

200

400

Time (sec)

Current (nA)

r

Current (nA)

Current (nA)

1.6

2 0x 100 x 200 x 300 x 400 x

1 0 0

200

400

Time (sec)

r = 2.9 mm

Figure 4.33: Photoresponse of a PbSe-MoS2 photodetector upon (a) inward bending up to a maximum bending radius of 3.8 mm and upon (b) outward bending up to a maximum bending radius of 2.9 mm. Inward bending results in a photocurrent increase, while outward bending leads to a photocurrent decrease. (c) Excellent photoresponse stability of the PbSe-MoS2 photodetector upon repeated bending (200 times outward bending, r = 4.4 mm; then 200 times inward bending, r = 4.3 mm). The ON-OFF switching time of the light source was 30 s, Vbias = 10 V. (Figure adapted from Schornbaum et al.[510] Copyright 2014 John Wiley & Sons, Inc.)

Both experiments, i.e., measuring the photoresponse under different bending radii and testing a high number of bending cycles, show the excellent mechanical stability of the in situ synthesized and subsequently solution-processed PbSe-MoS2 hybrid material. The results qualify PbSe-MoS2 hybrids for application in flexible and lightweight photodetectors. 120

4.5 Conclusion of Chapter 4

4.4.9 Summary The previous section introduced an air-stable and solution-processable PbSe-MoS2 photodetector. The in situ synthesized PbSe-MoS2 hybrid material was used to fabricate NIR sensitive photodetectors via a simple drop-casting process. The whole fabrication procedure was solution-based and no high-temperature steps were necessary. Therefore, photodetectors were made on rigid glass substrates as well as on flexible PET substrates. The devices were stable in ambient air, even after repeated illumination (∼ 50 times) and weeks of unprotected storage. There is no need to encapsulate the photodetectors to avoid oxidation of the PbSe QDs. This stability is the result of the ligand protection of the PbSe QDs at the outer surface, while there are no ligands at the PbSe-MoS2 interface. The direct interface enables a fast exciton dissociation under an applied bias, resulting in a clear and stable photoresponse upon illumination with NIR light (λ > 1200 nm). The 1200 nm cut-off filter was used to assure that the light absorption only happens through the PbSe QDs, while the charge carriers are transported within MoS2 nanoflakes. Finally, flexible devices demonstrate an excellent mechanical stability upon repeated bending and small bending radii.

4.5 Conclusion of Chapter 4 In order to use QDs in future applications, the overall charge carrier properties in QD-based devices still require improvement. Therefore, QDs were combined with high charge carrier mobility materials in order to make use of both, the efficient and tunable light absorption properties of QDs and the high charge carrier mobilities of the second material. A facile in situ hot-injection synthesis was developed to attach PbSe QDs non-covalently on different high-mobility nanomaterials. Hybrids of PbSe QDs with SWNTs, FLG, and multilayer MoS2 and WS2 were produced, showing the general feasibility of the synthesis technique. It was found that the nucleation of Pb2+ ions on the second material is crucial to achieve a dense coverage of the charge transporting material with PbSe QDs. A detailed characterization of the hybrids revealed a strong influence of the second material on the shape and the orientation of the QDs. PbSe QDs grow around the non-pretreated SWNTs, using the bundles as growth templates, thus leading to half ring shaped QDs. Within PbSe-MoS2 hybrids, PbSe QDs grow epitaxially on the MoS2 nanoflakes, resulting in a preferred orientation of the PbSe QDs on the MoS2 nanoflakes with three

121

4 Colloidal PbSe Quantum Dot Hybrid Materials equivalent orientation variants. Moreover it was shown that by choosing the right coupling strategy, a direct, linker-free interface between the QDs and the high charge carrier mobility material (e.g. SWNTs, MoS2 ) can be achieved. Finally, the characterization of photodetectors with PbSe-MoS2 hybrids revealed a clear NIR photoresponse, demonstrating an efficient charge separation at the PbSe-MoS2 interface. The photodetectors were long-time air-stable without encapsulation. This behavior is uncommon for devices based on PbX QDs (X = S, Se, Te) and attributed to the presence of protecting oleic acid ligands on the outer surface of the QDs. Moreover, flexible devices withstand small bending radii and repeated bending cycles, thus rendering them promising for future applications in low-cost and flexible NIR photodetectors. Although photodetectors of in situ synthesized PbSe-MoS2 hybrids are superior to many other QD or QD-hybrid photodetectors and already show very promising results, photocurrent generation can be further improved by various optimization processes. For example, a different electrode geometry with shorter electrode distances could be chosen, to increase electron extracting efficiencies. Moreover, it might be worth to take some efforts to achieve a more uniform coverage of the electrodes with the PbSe-MoS2 hybrids.

122

5 Quantum Dot Light-emitting Field-effect Transistors

This chapter demonstrates the fabrication and characterization of the first quantum dot light-emitting field-effect transistor. Ambipolar PbS quantum dot light-emitting field-effect transistors are realized based on ligand-exchanged QD thin-films with the help of electrolyte-gating. Characterization of the emission efficiencies revealed a significant enhancement of external quantum efficiencies, photoluminescence intensity and photoluminescence average lifetime with charge carrier accumulation. It is proposed that the increased emission efficiency at high charge carrier densities is a result of an effective deactivation of non-radiative decay channels and a dominant trion emission.

5 Quantum Dot Light-emitting Field-effect Transistors

5.1 Introduction Semiconductor quantum dots (QDs) are ideal candidates for solution-processed light emitting-devices, like light-emitting diodes (LEDs).[420,421,434,437] They exhibit a size-tunable emission wavelength, a narrow emission peak, and high photoluminescence (PL) quantum yields (QYs).[183] Especially in the near-infrared (NIR) regime, where efficient emitters are rare, NIR active PbS QDs represent an attractive alternative to commercially available NIR LEDs (e.g., InGaAs). The emission wavelength of PbS QDs can be freely tuned between 800-1800 nm.[431,437] Despite of the promising properties of QDs in solution, emission efficiencies of optoelectronic devices based on QD thin-films still require improvement. In order to optimize device performance, it is essential to understand in detail charge transport and recombination dynamics in QD thin-films. Many optoelectronic devices operate at high charge carrier densities, that is why the following experiments pay special attention to this condition. Ambipolar light-emitting field-effect transistors (LEFETs) are ideal devices to simultaneously investigate charge transport behavior as well as recombination and emission dynamics. LEFETs combine the switching properties of transistors with the light-emitting properties of LEDs. Additionally, LEFETs achieve very high charge carrier densities, which are orders of magnitude higher than in LEDs.[471] Under certain biasing conditions, both, electrons and holes are accumulated simultaneously within the transistor channel. The opposite charge carriers can subsequently radiatively recombine, leading to light emission from a narrow line. This confined recombination zone can be moved through the channel by varying the applied gate and/or source-drain voltage. If electron-hole recombination takes place in the middle of the channel far away from the electrodes, electrons and holes are perfectly balanced leading to the recombination of all charge carriers thus verifying true ambipolar charge transport (for more details on the operation of LEFETs and QD-based field-effect transistors (FETs) see Chapter 2, Section 2.5.2). However, LEFETs were so far only realized with bulk,[487,492] two-dimensional,[493,494] and one-dimensional materials,[495,496] but no LEFET based on zero-dimensional semiconductors were yet reported. This might be either due to an insufficient ambipolar charge transport in zero-dimensional materials or due to strong Auger quenching. PbS QDs seem to be the perfect material to fabricate the first LEFET based on a zero-dimensional material. They exhibit a nearly identical and large Bohr radius for electrons and holes (10 nm), this is why ambipolar charge transport, which is a basic requirement for the realization of LEFETs, is possible.[28]

124

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors This chapter demonstrates the fabrication and characterization of the first QD LEFET. PbS QD LEFETs are realized with the help of electrolyte-gating, which means that the conventional dielectric is replaced with an iongel. Electrolyte-gating helps to efficiently compensate trap states and dopants and thus enables access to the intrinsic ambipolar behavior of PbS QDs (see Chapter 2, Section 2.5.2 for details on electrolyte-gating). It is possible to detect NIR light emission (i.e., electroluminescence (EL)) from the channel region, due to the high charge carrier densities induced through electrolyte-gating.[478,482,497] Moreover, the freely tunable emission wavelength of these LEFETs is demonstrated by the fabrication of devices with three different sized PbS QDs. Subsequently, the voltage-dependence of EL and PL is investigated in detail via steady-state and time-resolved characterization techniques. These investigations provide better insights in the charge transport behavior, electron and hole accumulation as well as recombination and emission dynamics in PbS QD thin-film solids. Time-resolved PL experiments were performed together with Yuriy Zakharko.

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors 5.2.1 PbS Quantum Dots PbS QDs for all standard devices were synthesized by the hot-injection technique, using lead acetate and bis(trimethylsilyl) sulfide (TMS) as precursors, oleic acid as coordinating surfactant, and 1-octadecene as non-coordinating solvent (for details on the synthesis see Experimental Part, Section 3.2.2). The so-prepared QDs were dispersed in octane and stored in a nitrogen glovebox. PbS QDs of various sizes were synthesized, in order to achieve different emission wavelengths. Subsequently, the QD diameters were graphically extracted from bright field transmission electron microscopy (BF-TEM) images (Figure 5.1(a)) and cross-checked with data obtained from the corresponding absorption spectra (Figure 5.1(b)). The position of the first absorption maximum (λAbs = 1230 nm for PbS 1, λAbs = 1303 nm for PbS 2, and λAbs = 1407 nm for PbS 3) enabled the calculation of the PbS QD diameters (for calculation see Experimental Part, Section 3.4.1), yielding average diameters of 4.3 nm for PbS 1, 4.6 nm for PbS 2, and 5.1 nm for PbS 3. Additionally to the first synthesis route, PbS QDs were also synthesized with different precursors, namely lead chloride, elemental sulfur, and oleylamine. Oleyl-

125

(a)

PbS 1

20 nm

PbS 2

20 nm

PbS 3

20 nm

(b)

Normalized Absorbance

5 Quantum Dot Light-emitting Field-effect Transistors

1.0

0.5

0.0 1000

PbS 1 PbS 2 PbS 3 1200

1400

1600

Wavelength (nm)

Figure 5.1: (a) BF-TEM images of all three PbS QD batches (using lead acetate and TMS as precursors) and (b) corresponding normalized absorption spectra with absorption maxima at 1230 nm for PbS 1 (“green”), 1303 nm for PbS 2 (“red”), and 1407 nm for PbS 3 (“blue”), yielding average diameters of 4.3 nm for PbS 1, 4.6 nm for PbS 2, and 5.1 nm for PbS 3. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

amine simultaneously acts as solvent and coordinating surfactant. Figure 5.2 shows that these PbS QDs exhibit a very narrow size distribution with a fullwidth half maximum (FWHM) of around 95 nm and a well-pronounced second absorption peak. The absorption spectrum clearly demonstrates the narrower QD size-distribution of these PbS QDs compared to the QDs obtained by the first synthesis route. Moreover, these PbS QDs have ultrahigh PL QY reaching 40 %. Possible explanations for this are either the good passivation of trap states by the halogen anions originating from the lead precursor PbCl2 and/or the complete Pb-termination of the QD surface due to the large initial excess of lead (Pb:S 1:24; Pb is less prone to oxidation than S).[204,206] The high degree of monodispersity leading to a narrow emission linewidth in combination with the high PL QY renders these PbS QDs on first sight ideal for light-emitting PbS QD devices. However, several difficulties in using these QDs in LEFETs appeared: Due to the large excess of lead, a lot of unconsumed PbCl2 is left at the end of the synthesis, which is wasteful and results in the need for time-consuming and laborious separation steps to isolate the PbS QDs. Additionally, due to the large excess of lead precursor, it is difficult to upscale the PbS QD synthesis in order to get enough PbS QDs to perform comparative studies. Thus, if not stated otherwise, PbS QDs synthesized according to the first synthesis route, using lead acetate and TMS as precursors, were used for all following studies.

5.2.2 Quantum Dot Thin-films PbS QD thin-films were prepared by a layer-by-layer (LBL) spin-coating process. To obtain a smooth QD film, the PbS QD dispersion was filtered through a 0.2 µm

126

Normalized Intensity

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors

0.8

0.4

0.0 800

1000

1200

1400

1600

Wavelength (nm) Figure 5.2: Normalized absorption spectrum of PbS QDs (synthesized with lead chloride and elemental sulfur as precursors), showing an absorption maximum at 1414 nm and a narrow FWHM of 95 nm.

syringe filter thus avoiding any large aggregates. After each spin-coated PbS QD layer, a ligand exchange was performed from the long, insulating ligand oleic acid to the much shorter ligand 3-mercaptopropionic acid (MPA). This ligand exchange is necessary to enable charge carrier transport through the QD solid and thus realize measurable thin-film transistor characteristics. Moreover, exchanging ligands transforms the QD surface from nonpolar to polar. This polarity change enables a LBL film formation process, because it prevents dissolving the already deposited QD layers when applying another nonpolar PbS QD solution to form the next layer. All LEFETs were fabricated with PbS QD thin-films of 5-6 layers.

Normalized Intensity

The first excitonic absorption peaks of PbS QD thin-films are redshifted and broadened compared to the absorption peaks of the corresponding PbS QD dispersions (shown exemplarily for PbS 1, in Figure 5.3). This observation can be ascribed to a change in the QD dielectric environment through polarization effects. Moreover, in QD thin-films not all surface trap states can be passivated by an excess of ligands, in contrast to QD solutions.[90,120,203,538]

1.0

0.5

0.0 800

1000

1200

1400

Wavelength (nm) Figure 5.3: Normalized absorption spectra of PbS 1 QDs in dispersion (“blue”) and in a single-layer thin-film (“red”). The thin-film absorption is redshifted and broadened compared to the absorption in dispersion.

127

5 Quantum Dot Light-emitting Field-effect Transistors The position of the emission wavelength and the emission efficiency are two important characteristics of light-emitting devices. Therefore, the PL of PbS QD dispersions and of QD thin-films was investigated. Figure 5.4 depicts PL spectra of single-layer PbS QD thin-films capped with oleic acid (“purple”) and PL spectra of few-layer PbS QD thin-films after a ligand exchange from oleic acid to MPA (“cyan”). The emission peaks of the thin-films with MPA show a redshift of around ∼ 100-150 nm (depending on the PbS QD batch) compared to the emission peaks of thin-films with oleic acid. PbS 1

0.5

0.0 1200

1400

1600

Wavelength (nm)

PbS 3 (c)

1.0

Norm. Intensity

(b)

1.0

Norm. Intensity

Norm. Intensity

(a)

PbS 2

0.5

0.0 1200

1400

1600

Wavelength (nm)

1.0

0.5

0.0 1200

1400

1600

Wavelength (nm)

Figure 5.4: Normalized PL spectra of single-layer PbS QD thin-films still capped with oleic acid (“purple”) and of 5 LBL PbS QD thin-films after a ligand exchange from oleic acid to MPA (“cyan”) for (a) PbS 1, (b) PbS 2, and (c) PbS 3. Thin-films with MPA capped PbS QDs exhibit a redshift compared to thin-films with oleic acid capped PbS QDs. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

This redshift comes along with a substantial decrease in the PL QY. The PL QY of PbS QDs in dispersion yields relatively high values of around 12 %, while it decreases one order of magnitude to 1.5-7 % for single-layer PbS QD thin-films, where QDs are still capped with oleic acid. The quenching results from a combination of two different effects. First, the dipole-dipole coupling between individual QDs is stronger in films compared to solutions, due to a smaller distance of the QDs in thin-films than in solution. Second, the substrate provides additional PL quenching sites, leading to a reduction of the PL QY. When exchanging the oleic acid ligands with MPA, the PL QY further decreases around two orders of magnitude to low values of 0.05-0.07 % (see Table 5.1). This drastic decrease in PL QY is explained by a reduction of the interparticle spacing through the ligand exchange. While a large interparticle spacing facilitates a radiative recombination of excitons leading to high PL QYs, a small interdot spacing favors charge separation and subsequent transport resulting in a high charge carrier mobility.[2,437] In order to fabricate QD LEFETs, high charge carrier mobilities as well as high PL QYs are desirable. This is why the PbS interparticle spacing length has to be adjusted

128

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors in a way to balance both processes in the most efficient way. However, a certain charge transport over long distances is necessary to achieve currents high enough to obtain measurable EL from the transistor channel. Therefore, despite of the drastic PL quenching properties of MPA, this short capping ligand had to be used to form QD thin-films for LEFETs. QY [%]

PbS 1

PbS 2

PbS 3

Solution (Ligand: Oleic acid)

12

11

13

Single layer (Ligand: Oleic acid)

7

4

1.5

5LBL (Ligand: MPA)

0.05

0.07

0.07

Table 5.1: PL QYs of PbS 1, PbS 2, and PbS 3 in solution, in single-layer thin-films and in 5 LBL thin-films. The values demonstrate a decrease in PL QY from solution to thin-film and an even further decrease upon a ligand exchange from oleic acid to MPA.

The PbS QD thin-film morphology was examined via scanning electron microscopy (SEM) (Figure 5.5(a)). The SEM top-view image of a 5 LBL ligandexchanged PbS QD film reveals many cracks, resulting from the reduction of the interparticle spacing as a consequence of the ligand exchange process.[123] However, the corresponding SEM cross-section images clearly reveal that none of these cracks runs through the entire film (Figure 5.5(b) and (c)), why a continuous charge carrier transport is guaranteed. Moreover, side-view investigations show a homogeneous QD thin-film, with an average thickness of 220-240 nm. This homogeneity is achieved through the LBL spin-coating process, which assures that all cracks are filled through several spin-coating cycles. (a)

Top-view

(c)

Cross-section

2 µm

(b)

Protective layer

100 nm

PbS

400 nm

Si

Figure 5.5: (a) Top-view and (b), (c) cross-section SEM images of a 5 LBL ligand-exchanged PbS QD thin-film. The top-view image in (a) reveals surface cracks, which originate from an interparticle distance decrease upon ligand exchange. The cross-section images in (b) and (c) show the homogeneity of the PbS QD thin-film, with a layer thickness of around 230 nm. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

Film homogeneity could be improved through a post film formation annealing step for 30 min at 120 ◦ C. The top-view SEM image (Figure 5.6(a)) of the annealed

129

5 Quantum Dot Light-emitting Field-effect Transistors

1 µm

(c) 0.8

0.4

0.0 1200

Drain Current (A)

(b)

Norm. EL PL

(a)

Norm. Absorbance

sample shows a much smoother film with less cracks than the non-annealed film in Figure 5.5. However, annealing has some essential drawbacks. It broadens the PbS exciton PL peak and it redshifts the peak beyond the detector limit of 1600 nm. The same effect can be observed for the EL peak (Figure 5.6(b)). It is assumed that the observed effects result from a ligand loss upon annealing, which results in a sintering of the uncapped QDs. Therefore, despite of slightly higher drain currents in annealed QD thin-films compared to non-annealed films (Figure 5.6(c)), an annealing step cannot be applied to cure cracks. Based on these findings, all following experiments are conducted with non-annealed PbS QD thin-films.

1600

Wavelength (nm)

1E-5 1E-6 1E-7 1E-8 -1

0

1

2

Gate Voltage (V)

Figure 5.6: (a) Top-view SEM image and (b) normalized absorption (“black”), PL (“red”), and EL (“blue”) spectra of a 6 LBL PbS QD thin-film annealed at 120 ◦ C for 30 min. Annealing removes the cracks but at the same time redshifts the PL and EL peak beyond the detector limit (cut-off at 1600 nm). (c) Transfer characteristics of a FET with the same PbS QD thin-film before (“black”) and after (“red”) annealing, showing a drain current increase upon annealing. ((a) and (b) adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

5.2.3 Electrolyte-gated Thin-film Transistors All standard thin-film transistors were realized with electrolyte-gating, using an ionic liquid or iongel as dielectric in order to ensure high charge carrier densities (for alternative gating strategies see Section 5.2.6). In a first attempt, a polydimethylsiloxane (PDMS) boat was put on top of the PbS QD layer and subsequently this boat was filled with the ionic liquid 1-ethyl-3-methyl-imidazolium tris(pentafluoroethyl)-trifluorophosphate ([EMIM][FAP]). The PDMS boat is necessary to keep the ionic liquid in the confined region on top of the channel area. In this way a spreading of the ionic liquid onto the contact pads was avoided. A platinum wire was put into the ionic liquid and used as gate electrode. Figure 5.7 demonstrates the well-pronounced on- and off-switching ability as well as the clear ambipolar charge transport behavior of the so-fabricated thin-film transistors, with high electron and hole currents. However, due to the ionic liquid,

130

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors

Drain Current (A)

these devices cannot be encapsulated. PbS QDs are prone to oxidization, thus, when PbS QD thin-films were measured outside a glovebox without encapsulation, the QDs oxidized, which resulted in a strong p-doping of the film and a loss of ambipolar transistor characteristics. As a result, electrons and holes are no longer simultaneously present in the transistor channel, and thus no EL could be detected. Therefore, despite of the excellent ambipolar transistor characteristics inside a nitrogen glovebox, the ionic liquid gated devices could not be used to generate light emission. 10-4 10-5 10-6 10-7 0

1

2

Gate Voltage (V) Figure 5.7: Transfer characteristics of a PbS QD FET gated with the ionic liquid [EMIM][FAP].

In order to solve this problem, the ionic liquid was replaced with an iongel. The iongel was made of the ionic liquid [EMIM][FAP] and the polymer poly(vinylidene fluoride-cohexafluoropropylene) (P(VDF-HFP)), both dissolved in acetone. Right after spin-coating the iongel on top of the PbS QD thin-film, acetone evaporates and the liquid transforms into a solid gel. In order to facilitate encapsulation and avoid further process steps, a side-gate geometry was chosen to characterize the transistor (see Figure 3.2). Figure 5.8 displays the output and transfer characteristics of three thin-film transistors, fabricated with the different sized PbS QDs (PbS 1, PbS 2, PbS 3). All three devices exhibit a V-shaped transfer curve indicating ambipolar transistors. This V-shape is generated by keeping the drain current (Vd ) constant while simultaneously sweeping the gate voltage (Vg ). For low positive Vg the transistor operates in an unipolar electron accumulation regime, which changes into an ambipolar operation under low negative Vg and finally turns into an unipolar hole accumulation regime at high negative Vg . The forward and backward drain currents exhibit a large hysteresis, which appears in drain current differences for forward and backward scan. This large hysteresis can be observed for all three thin-film transistors, independent of the PbS QD size. Such a hysteresis is typical for transistors based on lead chalcogenide QDs.[117] It probably results from charge carrier trapping in numerous shallow and deep trap states.[117,155,201,497] Figure

131

5 Quantum Dot Light-emitting Field-effect Transistors PbS 1

PbS 2

Vg = +/- 0.5 V Vg = +/- 1.0 V Vg = +/- 1.5 V

-0.2

Vg = +/- 2.0 V

-1

0

1

Drain Voltage (V)

Drain Current (A)

(d)

0.0 Vg = 0 V Vg = +/- 0.5 V

-0.5

Vg = +/- 1.0 V Vg = +/- 1.5 V

-1.0

Vg = +/- 2.0 V

-1

0

Drain Voltage (V)

(e) N2 Glovebox 1E-3 1E-5 Vd = -0.4 V

Vd = -0.6 V

1E-7

Vd = -0.8 V Vd = -1.0 V

-2

-1

0

1

Gate Voltage (V)

1E-3 1E-5

Vg = +/- 0.5 V

-0.2

Vg = +/- 1.0 V Vg = +/- 1.5 V

-0.4

Vd = -0.6 V

1E-7

Vd = -0.8 V

Vg = +/- 2.0 V

-1

0

Vd = 0.6 V Vd = 0.8 V Vd = 1.0 V

2

Gate Voltage (V)

1

Drain Voltage (V) N2 Glovebox

Vd = -1.0 V

-2

-1

0

1

Gate Voltage (V)

Vd = -0.4 V Vd = -0.6 V

1E-7

Encapsulation

Drain Current (A)

Vd = 0.4 V

1

Vg = 0 V

Vd = -0.8 V Vd = -1.0 V

-2

-1

0

1

Gate Voltage (V) Encapsulation

(i) Air

Air

1E-6

0

0.0

1E-5

Vd = -0.4 V

Air 1E-4

-1

0.2

1E-3

(h)

1E-8

0.4

(f) N2 Glovebox

Encapsulation (g)

Drain Current (A)

1

Drain Current (mA)

Vg = 0 V

0.5

1E-3 1E-5 Vd = -0.4 V

Vd = -0.6 V

1E-7

Vd = -0.8 V Vd = -1.0 V

-2

-1

0

1

Gate Voltage (V)

Drain Current (A)

0.0

(c)

Drain Current (A)

0.2

Drain Current (mA)

(b)

Drain Current (A)

Drain Current (mA)

(a)

PbS 3

1E-4 1E-6

Vd = -0.4 V Vd = -0.6 V

1E-8

Vd = -0.8 V Vd = -1.0 V

-2

-1

0

1

2

Gate Voltage (V)

Figure 5.8: Output characteristics of FETs fabricated with (a) PbS 1, (b) PbS 2, (c) PbS 3 and (d),(e),(f) corresponding transfer characteristics measured in a nitrogen glovebox before encapsulation and (g),(h),(i) in air after encapsulation. The ambipolar behavior of the FETs are maintained in air, demonstrating an effective protection of the PbS QD thin-films against oxidation. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

5.8(d)-(i) depicts the influence of varying drain voltages at a constant gate voltage. The transfer curve shifts for different drain voltages, due to a change in the potential difference between drain and gate electrode. This drain current shift is typical for ambipolar transistors.[471] Moreover, electron and hole field-effect mobilities of all three thin-film transistors were extracted from the transconductance at low drain voltages (Vd ). Although in this study EL characteristics are more important than absolute values for charge carrier mobilities, mobilities were calculated to roughly quantify the devices. Electron filed-effect mobilities range between 0.04 and

132

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors 0.06 cm2 V−1 s−1 , while hole field-effect mobilities are one order of magnitude lower, yielding values between 0.003 and 0.009 cm2 V−1 s−1 (see Table 5.2 for all values and Experimental Part, Section 3.5.2 for details on the calculation of field-effect mobilities). max. µlin [cm2/Vs] µlin (e-)

µlin (h+)

PbS 1

PbS 2

PbS 3

as-prepared

0.055

0.056

0.038

encapsulated

0.011

0.059

0.016

as-prepared

0.003

0.009

0.009

encapsulated

0.003

0.005

0.008

Table 5.2: Electron and hole field-effect mobilities before and after encapsulation of FETs fabricated with the three different-sized PbS QD batches.

Electrolyte-gated transistors were also fabricated with a second iongel, in which the ionic liquid [EMIM][FAP] was replaced with 1-ethyl-3-methyl-imidazolium bis(trifluormethylsulfonyl) imide ([EMIM][TFSI]). The [TFSI] anion is smaller and thus more mobile than the [FAP] anion, which should lead to a faster formation of the electric double layers (EDLs) and an increase in drain current for [EMIM][TFSI] gated transistors compared to [EMIM][FAP] gated transistors. The transfer curves (Figure 5.9) demonstrate only small effects upon changing the anion. There is a little increase in the electron current, but a decrease in the hole current. Before the change of the iongel, the hole current was already lower than the electron current. Thus, in terms of EL it is important to increase the hole current, in order to bring electron and hole currents to the same values. Since the difference between electron and hole current even increased by using [EMIM][TFSI], electrolyte-gating was always realized with a [EMIM][FAP]-based iongel in all following experiments.

5.2.4 Electroluminescence LEFETs were encapsulated with a small piece of glass and an UV-hardened epoxy, in order to be able to characterize their light-emitting properties in air. Figure 5.8(g)-(i) shows the transfer curves of all three encapsulated PbS QD thinfilm transistors measured outside a glovebox under ambient air conditions. The curves still exhibit their characteristic V-shape, which unambiguously identifies the preservation of the ambipolar properties. This result shows the effective protection of the PbS QD thin-film against oxidation after encapsulation. Fieldeffect mobilities are hardly affected by encapsulation, only electron field-effect mobilities are slightly reduced (Table 5.2). The valley in the V-shaped transistor

133

5 Quantum Dot Light-emitting Field-effect Transistors (b) [EMIM][FAP]

1E-4 1E-5 1E-6 Vd = 0.8 V

1E-7

Vd = 0.9 V Vd = 1.0 V

1E-8 0

1

2

Gate Voltage (V)

Drain Current (A)

Drain Current (A)

(a) [EMIM][TFSI]

1E-4 1E-5 1E-6 1E-7

Vd = 0.8 V

1E-8

Vd = 1.0 V

Vd = 0.9 V

0

1

2

Gate Voltage (V)

Figure 5.9: Transfer characteristics of PbS QD FETs gated with (a) [EMIM][TFSI] and (b) [EMIM][FAP], showing similar characteristics for both iongels.

curve (marked in green in Figure 2.17) represents the ambipolar regime, where electrons and holes are simultaneously present. As depicted in Figure 2.19 in Chapter 2, these opposite charge carriers can recombine resulting in light emission from a confined region within the channel. In order to characterize EL properties, a movie was recorded over different gate voltages, while the drain voltage was kept constant. Figure 5.10(c) demonstrates four selected images of this movie, showing that the fabricated transistor exhibits EL in the NIR spectral region. Moreover, the images clearly indicate, that the position of the light emission can directly be controlled by the applied voltages. In this particular example, the drain voltage (Vd ) was kept at a constant value of -1.2 V, while the gate voltage was swept from +1.3 V to -2.5 V and back to +1.3 V. The recombination zone and thus the light emission was moved from the source electrode, through the channel to the drain electrode and vice versa. It is possible to detect light emission from the middle of the channel far away from the electrodes. Therefore, these experiments are unambiguous proof for a truly ambipolar electron-hole transport with an associated exciton formation by electron-hole recombination within the PbS QD thin-film. The present PbS QD thin-film transistors are the first LEFETs, which were ever fabricated with a zero-dimensional material. Independent of the position of the light emission, the emission zone appears as a very uniform line rather than single points or line fragments. This observation corroborates a uniform charge injection and charge transport, resulting from a very homogeneous PbS QD layer. These findings are in agreement with previous SEM investigations. The width of the emission zone is determined to be around

134

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors (c)

1

-2.0

channel

Recombination zone

1500

750.0

0.000

0

2

4

Distance (µm)

6

h+ source

2250

Norm. Intensity

-1.8

e- source

-1.6

3000

Vd = -1.2 V

Peak Intensity (a.u.)

-1.2

5 µm

Vg ≈ -2.5 V

Vg ≈ -1.6

0

e- source

ambipolar -1 0

(d) h+ source

Vg (V)

Vd = -1.0 V

electrode

Vd = -0.4 V

1E-7 hole transport

Gate Voltage (V) -1.4

Vg ≈ -2.5 V

electron transport

1E-5

-2

(b)

Vg ≈ -1.6 V

1E-3

electrode

Drain Current (A)

(a)

1

2

3

4

5

Distance (µm)

Figure 5.10: (a) Transfer characteristics of an iongel-gated PbS QD LEFET, indicating the hole and electron transport regime as well as the ambipolar region. (b) Illustration of the position (distance from hole injecting electrode) and intensity of an EL peak of a PbS QD LEFET dependent on the gate voltage (Vd = -1.2 V, integration time 10 s). (c) Four representative false-color NIR images (wavelength range 800-1600 nm, integration time 5 s) of a movie recording the recombination zone (i.e., light emission) of a PbS QD LEFET dependent on the gate voltage (Vd = -1 V; Vg was changed from +1.3 V to -2.5 V and back to +1.3 V; integration time per frame = 5 s) and (d) corresponding spectra of the position of the recombination zone (i.e., light emission) within the channel. The position can directly be controlled by the applied voltages. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

2 µm. However, it is assumed that the actual emission is even narrower and that the measured line width is probably limited by the resolution of the optical setup and not by the inherent emission properties of the QDs. It can be seen that the emission intensity is much higher at the hole source than at the electron source. This correlates with the drain current intensities, where the electron current is higher than the hole current resulting from a higher electron field-effect mobility compared to a hole field-effect mobility. As an alternative to the LEFETs described above, which were fabricated with PbS QDs synthesized with lead acetate and TMS precursors, transistors were also fabricated with PbS QDs synthesized with lead chloride and elemental sulfur. Figure 5.11(a) demonstrates that these transistors are also slightly ambipolar, but exhibit a predominant n-doping. As introduced in Section 5.2.1, these PbS QDs were synthesized with a huge excess of lead chloride, resulting in a complete Pb termination of the QD surface. While a S surface termination of the QDs

135

5 Quantum Dot Light-emitting Field-effect Transistors

1E-6

Vd = 0.6 V Vd = 0.8 V Vd = 1.0 V

1E-8

Vd = 1.2 V

-1

0

1

2

Gate Voltage (V)

0.5

0.0 1000 1200 1400 1600

channel

1E-4

(c) NIR-light emission 1.0

50 µm

electrode

(b)

Norm. EL PL

Drain Current (A)

(a)

Norm. Absorbance

would induce p-doping, a Pb terminated surface leads to n-doping.[117,506] This information from literature explains the trends seen in the transistor curve in Figure 5.11(a). In contrast to thin-films of PbS QDs synthesized with lead acetate and TMS, PbS QD thin-films produced with QDs synthesized with lead chloride and elemental sulfur showed no noticeable broadening or redshifting of the PL peak upon annealing to 120 ◦ C for 30 min (Figure 5.11(b)). Most probably this is caused by the good passivation of the QD surface with excess Pb and Cl ions. Figure 5.11(b) shows, that it is possible to detect EL from these devices with a very narrow emission peak exhibiting a FWHM of ∼ 140 nm. However, despite of annealing, the overall drain current and especially the hole current of these transistors were low. Therefore, EL was mainly confined to the hole injecting electrode and as a consequence the light emission zone could not be moved through the channel by changing the applied voltages (Figure 5.11(c)). This property classifies these QD LEFETs as unipolar rather than ambipolar devices. As long as not stated differently, all of the following experiments were performed with PbS QDs synthesized with lead acetate and TMS precursors.

Wavelength (nm)

Figure 5.11: (a) Transfer characteristics of an electrolyte-gated LEFET fabricated with PbS QDs, which were synthesized using lead chloride and elemental sulfur as precursors. (b) The corresponding normalized absorption (“black”), PL (“red”), and EL (“blue”) spectra all exhibit very narrow optical peaks and almost no Stokes shift. Despite of annealing the PbS QD thin-film to 120 ◦ C for 30 min, both emission peaks are only slightly redshifted. (c) False-color NIR image (wavelength range 800-1600 nm) showing bright EL. However, the EL cannot be moved through the channel, it is mainly confined to the hole injecting electrode (Vg = 2 V, Vd = -1 V, integration time 30 s). (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

Electroluminescence and Photoluminescence Spectra PbS QDs are especially useful for the fabrication of LEFETs, because their band gap and thus their EL and PL emission wavelength can be easily tuned to the desired wavelength. This property renders QDs advantageous to other materials with a fixed band gap and thus a fixed emission wavelength. In order to illustrate these unique properties, PbS QD LEFETs were fabricated with three different sized PbS QD batches. 136

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors Subsequently, EL and PL spectra of all three LEFETs were recorded. EL was measured by applying a bias to the gate and the drain electrode. Afterwards, without moving the device, PL was detected from the same region, by exciting the PbS QD thin-film with a 640 nm laser. Figure 5.12 shows that all three devices emit light at different wavelengths. The emission wavelengths are in agreement with the wavelengths measured previously with absorption spectroscopy (see Figure 5.1(b)). The results clearly show the versatile utilizability of the quantum confinement effect of PbS QDs for light-emitting devices, in order to tune the emission wavelength in an easy and controllable way. Figure 5.12 depicts, that PL and EL spectra exhibit the same shape, with a FWHM of around 200 nm and a small redshift of the EL of around 10 nm compared to PL. The fact that the shapes of both curves perfectly resemble each other indicates, that the recombination of electrons and holes, happens within the QDs and not at their surface. There is no change in the shape of the EL curve, even when the applied voltages and thus the current densities are increased. The width of the absorption band (Figure 5.1(b)) indicates that the QDs are not perfectly monodisperse. This implicates that each QD batch is composed of QDs with slightly different sizes. However, due to the large gating-effect, which results from electrolyte-gating, it is assumed that the charge transport happens through all QDs, independent of their size.

5.2.5 External Quantum Efficiency External quantum efficiencies (EQEs) of the devices were measured in order to compare the presented LEFETs to state-of-the-art QD LEDs. Moreover, EQE characteristics usually provide information on the charge carrier transport and the emission properties in PbS QD thin-films at high charge carrier densities. Therefore, determining EQEs helps to get a deep understanding of charge carrier transport in the presence of high charge carrier densities. Figure 5.13 depicts the EQE of a PbS QD LEFET at different drain voltages and thus different current densities. The maximum EQE was determined to be 0.002 % and was obtained for forward emission (see Experimental Part, Section 3.5.2 for details on the determination of the EQE). This value is similar to the EQE of MPA ligand-exchanged PbS QD LEDs. However, the EQE in comparable PbS QD LEDs decreases with high current densities,[437] while the EQE in the here presented PbS QD LEFETs constantly increases with current densities, even beyond 10 A/cm2 (see Experimental Part, Section 3.5.2 for the calculation of the current density). This different behavior might be caused by the device structure: LEFETs exhibit a planar structure with

137

5 Quantum Dot Light-emitting Field-effect Transistors

Normalized Intensity

Electroluminescence

Photoluminescence exc = 640 nm

1200

1400

1600

Wavelength (nm) Figure 5.12: Normalized EL (Vg = 0.1 V, Vd = -1.2 V, integration time 60 s) spectra and corresponding normalized PL (λexc = 640 nm, ∼10 mW/µm2 , integration time 20 s) spectra from the same spot within the channel region of three LEFETs with the different-sized PbS QD batches (PbS 1 = “green”, PbS 2 = “red”, PbS 3 = “blue”). EL and PL spectra exhibit the same shape, both with a FWHM of around 200 nm. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

a long channel length of 5 µm, while LEDs have a film thickness of around 200 nm. Thus, unlike to LEDs, in LEFETs a complete electron-hole recombination occurs, possibly explaining the different EQE characteristics. A further increase of the current density to achieve an EQE higher than 0.002 % was not possible due to the limited electrochemical window of the iongel. Usually additional charges decrease the QY,[539–541] therefore the observed increase in EQE with increasing current density is an interesting behavior and will be further investigated in Section 5.3.

EQE (%)

0.003

(b)

Vd = -1.0 V Vd = -1.2 V

EQE (%)

(a)

0.002 V = -1.4 V d 0.001

10-2

EQEmax ~ 0.002 % 10-3

0.000 -2

-1

0

Gate Voltage (V)

1

1

10

Current Density (A/cm2)

Figure 5.13: (a) EQEs of a PbS QD LEFET dependent on the gate voltage at different drain voltages. (b) EQE versus current density, showing a steady increase of EQE with current density up to a maximum value of around 0.002 %. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

138

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors However, it should be noted that Sun et al. increased the EQE of PbS QD LEDs to a maximum value of around 2 %, which is three orders of magnitude higher than the EQE in the current PbS QD LEFETs. The high EQE of 2 % was achieved with QDs covered with the ligand 8-mercaptooctanoic acid (MOA).[437] MOA exhibits eight CH2 groups, compared to MPA with only three CH2 groups, and thus leads to a larger interparticle distance than MPA. Therefore, the possibility for charge carrier recombination is increased at the expense of charge carrier separation and transport, resulting in an enhanced EQE. Based on these results, a PbS QD LEFET was fabricated with a 6-mercaptohexanoic acid (MHA) ligand-exchanged QD thin-film. MHA exhibits six CH2 groups and was used to increase radiative recombination of excitons while still maintaining good charge carrier transport characteristics. The intention was to achieve the perfect balance between a high current and a high EQE. Figure 5.14 shows the transistor characteristics of one PbS QD LEFET with MPA ligand-exchanged PbS QDs and a second LEFET with MHA ligand-exchanged PbS QDs. The drain current is reduced by using MHA, reflecting the longer interparticle spacing in MHA ligand-exchanged thin-films. Consequently, EL is weak and hardly detectable and thus MPA and not MHA ligand-exchanged PbS QD thin-films were used for all further experiments. (b) MHA

1E-4 1E-5 1E-6 Vd = 0.8 V

1E-7

Vd = 0.9 V

1E-8

Vd = 1.0 V

-1

0

1

Drain Current (A)

Drain Current (A)

(a) MPA

1E-4 1E-5 1E-6 Vd = 0.8 V

1E-7

Vd = 0.9 V

1E-8

2

Gate Voltage (V)

Vd = 1.0 V

-1

0

1

2

Gate Voltage (V)

Figure 5.14: Transfer characteristics of iongel-gated FETs fabricated with PbS QD thin-films with (a) MPA and (b) MHA ligands, showing a reduced drain current for FETs based on MHA ligand-exchanged QD thin-films compared to MPA ligand-exchanged QD thin-films.

5.2.6 Top-gated Thin-film Transistors In order to test if electrolyte-gating is the best choice for PbS QD LEFETs, topgated transistors were prepared, with either poly(methyl methacrylate) (PMMA) or HfOx as dielectric. First, LEFETs were fabricated by spin-coating PMMA

139

5 Quantum Dot Light-emitting Field-effect Transistors

1E-6

Drain Current (A)

(a)

(b)

1E-7 Vd =

1E-8

10 V 15 V 20 V 25 V 30 V

1E-9 1E-10

-10

0

10

20

30

Gate Voltage (V)

Drain Current (A)

on top of the ligand-exchanged PbS QD thin-film, followed by an evaporation of a silver gate electrode. Figure 5.15 depicts the transistor characteristics of two devices, showing a slightly ambipolar but mainly p-type behavior for PbS QDs synthesized with lead chloride and elemental sulfur (Figure 5.15(a)), while transistors fabricated with PbS QDs synthesized with lead acetate and TMS only exhibit hole transport and very little current modulation (Figure 5.15(b)). A basic requirement to achieve EL is an ambipolar charge transport. Therefore PMMA was not further considered for the fabrication of PbS QD LEFETs.

1E-6

Vd = 10 V 15 V 20 V 25 V 30 V

1E-7 -10

0

10

20

30

Gate Voltage (V)

Figure 5.15: Transfer characteristics of Ag top-gated PbS QD LEFETs with a PMMA dielectric. The QDs were synthesized using (a) lead chloride and elemental sulfur or (b) lead acetate and TMS.

In a following step it was examined if a thin HfOx dielectric layer, deposited via atomic layer deposition (ALD), is suitable for the fabrication of PbS QD LEFETs. It was expected that HfOx infills the QD film and thus protects the QDs from oxidation, avoiding subsequent encapsulation steps.[503] Indeed, transistor curves measured inside a nitrogen glovebox (Figure 5.16(a)) and in air (Figure 5.16(b)) exhibit the same characteristics, corroborating an effective protection of the PbS QDs against oxidation. However, transistors are only slightly ambipolar with a dominant n-type behavior, which leads to mainly unipolar FETs. Interestingly, despite of the high drain current EL is very weak, which might be due to a quenching caused by the HfOx . Moreover, the weak EL only occurs at the hole injection electrode and cannot be moved through the channel by changing the applied voltages (Figure 5.16(c)). Thus, fabricating transistors with ALD deposited HfOx provides a good possibility to protect PbS QDs from oxidation but further optimization needs to be done to increase currents in order to use this promising dielectric for QD LEFETs. However, this optimization process goes beyond the scope of this thesis.

140

5.2 PbS Quantum Dot Light-emitting Field-effect Transistors

1E-5

Vd = -1.0 V -5.0 V -6.0 V -7.0 V

1E-7 1E-9

-10 -5

0

5

10

Gate Voltage (V)

1E-3

(c)

Air

1E-5

Vd = -1.0 V -5.0 V -6.0 V -7.0 V

1E-7 1E-9

NIR-light emission channel

(b)

N2 Glovebox

-10 -5

0

5

10

Gate Voltage (V)

50 µm

electrode

1E-3

Drain Current (A)

Drain Current (A)

(a)

Figure 5.16: Transfer characteristics of a Ag top-gated PbS QD LEFET with a 50 nm HfOx dielectric characterized (a) inside a nitrogen glovebox and (b) in air without encapsulation. Both curves exhibit the same shape, corroborating an effective protection of the PbS QDs against oxidation. (c) False-color NIR image (wavelength range 800-1600 nm) showing weak EL. The EL is mainly confined to the hole injecting electrode and cannot be moved through the channel (Vg = 9 V, Vd = 6 V, integration time 5 s). (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

5.2.7 Summary This section discusses the fabrication of the first QD LEFETs. The presented devices are based on electrolyte-gated PbS QDs and show a wavelength tunable and position movable NIR emission. PbS QDs of different sizes were synthesized following a standard hot-injection recipe. The optical properties of all batches are different, however the absorption wavelengths lie within the NIR spectral range for all dots. PbS QD thin-films were achieved through a LBL spin-coating process in combination with a ligand exchange from oleic acid to MPA. The ligand exchange transforms the initially insulating layers into a conducting thin-film with the unavoidable trade-off of drastically decreasing the QD PL QY. In a next step, electrolyte-gated, ambipolar LEFETs were fabricated using the ligand-exchanged PbS QD thin-films. The transistors were encapsulated and thus the ambipolar behavior could be preserved in ambient air. NIR light-emission was detected from a confined region within the transistor channel, resulting from a recombination of electrons and holes. The emission wavelength was directly correlated to the PbS QD size. Moreover, the position of the emission zone could be moved through the channel by changing the applied gate and/or drain voltage. Due to the higher electron than hole mobility in the PbS QD thin-films, EL is more intense at the hole injecting electrode than at the electron injecting electrode. A characterization of the light emission efficiencies revealed an increase in EQE with increasing current densities. This unusual behavior will be studied in more detail in the following section.

141

5 Quantum Dot Light-emitting Field-effect Transistors

5.3 Charge Carrier Dynamics in PbS Quantum Dot LEFETs 5.3.1 Gate Voltage Dependence of Photoluminescence The PbS QD thin-films showed and unexpected increase in EQE with increasing charge carrier densities (see Figure 5.13). Usually doping quenches excitons and thus lowers the EQE,[539–541] this is why this contrary observation is very interesting. Several experiments were conducted in order to understand this behavior. First, PL spectra of the channel region were recorded under a constant laser excitation (λexc = 640 nm) and a gate bias. Additionally, a small drain voltage Vd of 10 mV was applied in order to be able to simultaneously monitor the conductivity of the QD thin-film. Figure 5.17(a),(b), and (c) clearly show that the drain current (“purple”) and the gate-voltage dependent PL intensity (“black”) follow the same trend for all three LEFETs. At the threshold voltage for electron and hole transport, the integrated PL intensity sharply increases about 2-3 times. The corresponding PL spectra in Figure 5.17(d),(e), and (f) illustrate that charge accumulation in PbS QD thin-films does not induce spectral changes. Although the applied drain voltage Vd is very small and a detectable generation of EL is highly unlikely, a control experiment with the same biasing conditions but without laser excitation was performed. This measurement excluded any EL generation. The described experiments are in principle electrochemical doping experiments, where the source/drain electrode can be considered as the working electrode, while the gate electrode represents the counter electrode.

5.3.2 Transient Photoluminescence Analysis The experiments in Section 5.3.1 reveal that the PL intensity is influenced by additional charges. However, it is not possible to draw conclusions on charge carrier dynamics. Therefore, PL lifetime measurements were performed, in order to investigate the influence of doping on the emission properties of PbS QD thinfilms. Different levels of electron and hole doping of the PbS QD thin-films were obtained by changing the applied gate voltages. In order to avoid multiexciton contributions, a low excitation power was used (see Experimental Part, Section 3.4.2 for calculations of the excitation power). Figure 5.18(a),(b), and (c) depict the PL intensity (“black”) and the PL average lifetime (“blue”) for all three PbS LEFETs (fabricated with PbS 1, PbS 2, and PbS 3) (see Experimental Part,

142

10-6

2000 10-7 1500 10-8 1000 -3

-2

-1

0

1

2

10-9 3

(d)

1200

10-7

200 150

10-8

100

10-9 -2

0

(e)

Vg =

50

10-8 0

2

Gate Voltage (V)

1600

(f)

Vg = -2.2 V -1.4 V -1.0 V 0.0 V 1.0 V 2.2 V

PL Intensity

10-7

Drain Current (A)

Integrated PL Intensity

PbS 3

100

-2

1400

Wavelength (nm) 10-6

150

-2.2 V -1.0 V 0.0 V 1.0 V 1.4 V 2.2 V

1200

2

Gate Voltage (V)

(c)

1600

PL Intensity

10-6

Drain Current (A)

Integrated PL Intensity

PbS 2

300 250

1400

Wavelength (nm)

Gate Voltage (V)

(b)

Vg = -2.2 V -1.4 V -1.0 V 0.0 V 1.0 V 1.4 V 2.2 V

PL Intensity

2500

Drain Current (A)

PbS 1

(a)

Integrated PL Intensity

5.3 Charge Carrier Dynamics in PbS Quantum Dot LEFETs

1200

1400

1600

Wavelength (nm)

Figure 5.17: Integrated PL intensity from the channel region (“black”) and corresponding drain current (“purple”) (Vd = -10 mV) of (a) PbS 1, (b) PbS 2, and (c) PbS 3 QD LEFETs under a constant laser excitation (λexc = 640 nm, ∼10 mW/µm2 ). Both curves follow the same trend for all three LEFETs. (d),(e),(f) Corresponding PL spectra for selected gate voltages. ((a),(b),(c), and (d) adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

Section 3.4.2 for the definition of average lifetime), while Figure 5.18(d),(e), and (f) show the corresponding PL transients for selected gate voltages. It is obvious that both curves have the same gate voltage dependence, which is also consistent with the characteristics of the integrated PL intensity. Both, PL intensity and PL average lifetime increase for electron and hole accumulation, which coincides with the threshold voltage for charge transport. There are two possible explanations that might describe PL intensity changes with charge accumulation. Either doping increases the number of QDs with a high PL QY, while it does not change PL QY values, or it increases the PL QY of the actual emitting QDs, but keeps the

143

5 Quantum Dot Light-emitting Field-effect Transistors

15

5

12

4

9

3

6 -3

-2

-1

0

1

2

2 3

1

(d) Norm. Intensity

6

Average Lifetime (ns)

PbS 1

(a)

Integrated PL Intensity 3 (x10 counts/s)

number of emitting QDs constant. Besides PL intensity changes there is also an influence on PL average lifetime. Such a change in PL average lifetime is usually induced by variations of the radiative and non-radiative recombination channels. IRF Vg = 2.5 V

Vg = 2.0 V Vg = 0.0 V

0.1

Vg = -1.5 V Vg = -2.5 V

0.01 0

15

6

10

4

5 -3

-2

-1

0

1

2

2 3

(e)

Vg = 2.0 V Vg = 0.0 V

0.1

Vg = -1.5 V Vg = -2.5 V

0.01 0

16

6 5

12

4 8 -3

3 -2

-1

0

1

2

Gate Voltage (V)

3

(f)

20

30 IRF Vg = 2.5 V

1

Norm. Intensity

7

Average Lifetime (ns)

Integrated PL Intensity 3 (x10 counts/s)

PbS 3

8 20

10

Time (ns)

Gate Voltage (V)

(c)

30 IRF Vg = 2.5 V

1

Norm. Intensity

8

20

20

Time (ns) Average Lifetime (ns)

PbS 2

(b)

Integrated PL Intensity 3 (x10 counts/s)

Gate Voltage (V)

10

Vg = 2.0 V Vg = 0.0 V

0.1

Vg = -1.5 V Vg = -2.5 V

0.01 0

10

20

30

Time (ns)

Figure 5.18: Integrated PL intensity (λexc = 785 nm, pulsed laser diode) dependent on the gate voltage (“black”) and corresponding average emission lifetime (“blue”) at different doping levels of (a) PbS 1, (b) PbS 2, and (c) PbS 3 QD LEFETs. (d),(e),(f) Corresponding normalized PL decays for selected gate voltages. IRF indicates the instrument response function. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

In addition to the excitons created by laser excitation, a large number of electrically induced charges are injected into the QD solid by applying a gate voltage. Through this combination of a neutral exciton with an extra charge a charged three-particle bound state is formed, which is called trion. Negative trions are excitons with an extra electron, while positive trions are excitons with an extra hole (see Figure 2.6).[111] The lifetime of trions is substantially shorter

144

5.3 Charge Carrier Dynamics in PbS Quantum Dot LEFETs than the lifetime of excitons due to their fast Auger decay rate (non-radiative losses) and the higher chance of a radiative decay (see Chapter 2, Section 2.1.2 for details on trions in QDs). Consequently, trions could exhibit a higher PL QY than excitons.[111,112,542] The experimentally measured changes in PL intensity and PL average lifetime with electrical doping cannot be explained by a simple trap state compensation model, but they can be explained with a model based on an increasing number of trions. In order to get better insights into the charge carrier dynamics of PbS QDs, analyses of the PL intensity decays were performed. From initial conditions to positive gate voltages (0 V < Vg < 2 V), there are no mobile charge carriers in the QD thin-film. In this gate voltage range, the PL intensity hardly changes and the PL decay is dominated by a fast decay component. Due to the low PL QY of the thin-films (∼ 0.05-0.07 %) and the long exciton radiative lifetimes (1-3 µs of PbS QDs in solution) (see Experimental Part, Section 3.4.2 for calculations), it can be stated that the fast decay component corresponds to a fast non-radiative decay channel. An additional slow decay component appears, if the gate voltage is increased above the threshold voltage for electron or hole transport (Vg > 2 V or 0 V > Vg > -2 V). At high negative gate voltages (Vg = -2.5 V) the slope of the decay curve increases, which indicates a decrease of the corresponding time constant (Figure 5.17). In order to interpret these trends, non-linear transient analysis was performed to separate the slow decay component from the fast decay components, with a method similar to the one applied by Saba et al.[115] First, the PL decay curve, which was obtained for initial conditions (Vg = 0 V) was subtracted from all other transient curves (Figure 5.19(a),(b),(c)). The resulting decay curves were fitted with a mono-exponential decay function, which enabled to identify lifetimes between 1-12 ns. These lifetimes correspond to the slow decay components emerging at high doping levels (positive or negative Vg ), because the initial decay curve was subtracted before fitting. The average lifetime at Vg = 0 V is 2-3 ns and thus lower than most of the lifetimes identified for the slow decay component. Consequently, these results are in line with the observed trends of an increase in PL average lifetime with doping. However, some values within the trend line are missing, due to a subtraction of the PL decay curve for Vg = 0 V (Figure 5.19(d),(e),(f)). In order to obtain these missing data points and to cross-check the reliability of the extracted values, a second method was applied to analyze the PL transients. This method is based on a tri-exponential fit of the PL decays in Figure 5.18(d),(e), and (f), yielding two fast and one slow component (for fit residuals see Figure 3.7 in Experimental Part, Section 3.4.2). Moreover, Figure 5.19(d),(e), and (f) illustrate that the lifetimes obtained by the two different 145

5 Quantum Dot Light-emitting Field-effect Transistors methods show an excellent correlation and therefore the reliability of the results is high. (d)

Vg = 2.0 V

10

Vg = -1.5 V ( = 4.55 ns) Vg = -2.5 V ( = 3.15 ns)

1 0

1 10

20

10

Lifetime (ns)

Vg = 2.5 V ( = 9.37 ns)

PL Intensity (counts/s)

PbS 1

(a)

8 6 4 -3

30

Vg = 2.5 V (= 12 ns)

(e) 12

Lifetime (ns)

PL Intensity (counts/s)

PbS 2

100

Vg = 2.0 V Vg = -1.5 V (= 4.02 ns)

10

Vg = -2.5 V (= 1.66 ns)

1 0

10

20

Vg = 2.0 V Vg = -1.5 V (= 3.88 ns) Vg = -2.5 V (= 1.11 ns)

1 0

10

20

Time (ns)

1

2

3

Mono-exponential fit Tri-exponential fit

-2

-1

0

1

2

3

Gate Voltage (V)

Vg = 2.5 V (= 5.59 ns)

10

0

4

-3

30

30

(f)

Lifetime (ns)

PL Intensity (counts/s)

PbS 3

100

-1

8

Time (ns) (c)

-2

Gate Voltage (V)

Time (ns) (b)

Mono-exponential fit Tri-exponential fit

6 4 2 -3

Mono-exponential fit Tri-exponential fit

-2

-1

0

1

2

3

Gate Voltage (V)

Figure 5.19: PL decay curves (λexc = 785 nm, pulsed laser diode) of (a) PbS 1, (b) PbS 2, and (c) PbS 3 QD LEFETs for different gate voltages after subtracting the PL decay curve for initial conditions (Vg = 0 V). PL lifetimes of the slow decay component of (d) PbS 1, (e) PbS 2, and (f) PbS 3 QD LEFETs resulting from a mono-exponential decay fit (“black”) and a tri-exponential fit (“red”) of the PL decays in Figure 5.18(d),(e),(f). (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

The decay lifetimes for hole doping differ from that of electron doping for all three QD devices. The lifetimes for the electron accumulation regime (Vg > 2 V) are longer than for the hole accumulation regime (Vg < 0 V). This experimental finding is unexpected and might be explained by different lifetimes of positive and negative trions. In order to strengthen this assumption, a comparison with existing literature reports is done. It is expected that the conduction band (CB) and valence band (VB) in PbS QDs are mirrored around the Fermi level, which would lead to

146

5.3 Charge Carrier Dynamics in PbS Quantum Dot LEFETs equal lifetimes of positive and negative trions. However, in order to get different positive and negative trion lifetimes in PbS QD thin-films, electrons and holes would have to be asymmetrically localized around the Fermi level. Such an asymmetric localization of electrons and holes is possible and might result from charged surface ligands or/and strong local fields induced by electrolyte gating. Moreover, there are also publications reporting that the density of states effective mass of holes and electrons is not the same (for PbS: m∗h /m∗e ≈ 0.6342).[543] Furthermore, trion lifetimes are usually deduced from biexciton Auger-decay lifetimes. Stewart et al. experimentally determined the biexciton Auger-decay lifetime of oleic acid covered PbS QDs to be less than 100 ps, which yields a trion lifetime of a few hundred ps.[112,544] The trion lifetimes in PbS QD thin-films of this study are determined to be 1-12 ns, which is orders of magnitude longer than the values by Stewart et al. All in all, literature reports dealing with trion lifetimes in QDs are rare, therefore no clear verification or rejection of the aforementioned assumption can be obtained.[111–113,115] Moreover, regarding emission efficiencies, most reports consider trions as “dark” or “gray” states.[111,112,542] But despite of these reports, the introduction of trions in the proposed model is necessary and reasonable. Nevertheless, the differences between literature reports and the here proposed trion-based model should not be neglected. Additional experiments could be done to get a more accurate overall picture. Not only the PL average lifetime but also the PL intensity reveals changes with doping density. The PL intensity increases with increasing electron and hole accumulation. In order to understand this doping induced intensity enhancement, the amplitudes of the exponents from the tri-exponential fit were analyzed. Therefore, normalized values of the slow components and of the sum of the two fast components were used (Figure 5.20). First, the slow component will be analyzed in more detail. This component represents the trion-related emission. For unbiased conditions (Vg = 0 V) the fraction of QDs exhibiting a slow decay component ranges between 1-3 %, which increases up to 35 % for both, electron and hole accumulation (Vg > 2 V and Vg < 0 V) (Figure 5.20). In contrast to the slow component, the sum of the fast components is barely influenced by doping. The fraction of the fast component only slightly decreases, while the lifetime of the fast component is fixed. These results indicate an increasing fraction of QDs with a slow decay component induced by electrical charge injection. Therefore, it is corroborated that the modulation of the number of emitting QDs is probably responsible for PL intensity variations. Considering all results obtained by transient PL analysis, it can be stated that the emission caused by the slow decaying component, i.e., trion emission, is more 147

5 Quantum Dot Light-emitting Field-effect Transistors PbS 1 Sum of fast components Slow component

Fraction (%)

Fraction (%)

Sum of fast components Slow component

10

-2

0

2

Gate Voltage (V)

10

1

(c)

Sum of fast components Slow component

100

100

100

1

(b)

PbS 3

Fraction (%)

(a)

PbS 2

-2

0

2

Gate Voltage (V)

10

1

-2

0

2

Gate Voltage (V)

Figure 5.20: Relative contribution of the sum of the two fast decay components (“black”) and the slow decay component (“red”) (all three extracted from the tri-exponential fit) to the overall light emission dependent on the gate voltage for (a) PbS 1, (b) PbS 2, and (c) PbS 3 QD LEFETs. Both dependencies show that only the slow component is responsible for PL intensity changes upon charge accumulation. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

intense than the emission caused by the fast decaying component. Consequently, in an unbiased QD thin-film (Vg = 0 V) only 1-3 % of all PbS QDs exhibit a trion-related component and thus, 97-99 % of all initial dots are inefficient emitters.

5.3.3 Proposed Model of Recombination Channels The transformation of many inefficient emitting QDs (“dark” QDs) into efficient emitting QDs (“bright” QDs) can be explained with a decrease of non-radiative losses through charge accumulation. In each dot in Figure 5.21 the Fermi level EF and the energetic levels for electron traps Et− and/or hole traps Et+ are marked. The Fermi level is shifted by applying a gate voltage, changing the relative position of EF to Et− or Et+ . This shift can cause an activation or deactivation of recombination centers, i.e., traps.[542] The same effect is also responsible for the switching properties of FETs, where charge carriers become mobile above the threshold voltage due to a complete filling of all trap states.[545] Electron or hole traps states, which are also called midgap states, can be induced by defects, ligands, and/or stoichiometry variations.[200,546–548] It is highly probable that the ligand exchange, which was performed during thin-film formation, induced defects (i.e., midgap states). Moreover, all PbS QDs exhibit some defects already right after synthesis and usually are non-stoichiometric (see Chapter 2, Section 2.1). Thus, it can be expected that midgap states are present in most of the processed PbS QDs. Based on these facts, the proposed model contains two subgroups of

148

5.3 Charge Carrier Dynamics in PbS Quantum Dot LEFETs PbS QDs. The first subgroup exhibits PbS QDs with an excess of electron traps (“blue” dots in Figure 5.21), while the second subgroup is composed of PbS QDs with an excess of hole traps (“red” dots in Figure 5.21). In some QDs the number of electron and hole traps might be equal, and consequently compensate each other (“green” dots in Figure 5.21). The deactivation of electron and hole trap states leads to an increase in the ratio of radiative to non-radiative decay rates. This results in more efficient emitting QDs and finally an overall enhanced PL. Recombination channels in QDs Electron accumulation Vg > 2 V

Hole accumulation

-EF Et-

-

γrXγnrX-

EF Et-

EtEF

γnr

+ --

Et-

EF

+ -

Et+

++ Et-

EtEF Et+

γnr

γnr

γrX+ γnrX+

γrX0 γnrX0

+ - -

Vg < 0 V

0 V < Vg < 2 V

Et+ EF

+ -

γnr

+ ++ -

EF Et+

+

γnr

EF Et+

+

γnr

Et+ EF

+ ++

Excess of electron traps Excess of hole traps Deactivated traps Figure 5.21: Scheme of the proposed model for the different recombination channels in PbS QDs. The middle column represents the initial conditions (0 V < Vg < 2 V), while the left column and the right column represent the electron accumulation (Vg > 2 V) and hole accumulation (Vg < 0 V) regime, respectively. The schematic blue dots indicate QDs with an excess of electron trap states (Et− ), while red dots mark QDs with an excess of hole trap states (Et+ ) (γnr is the effective trapping rate). The green-colored schematic dots represent the QDs, where all trap states are deactivated. The gate voltage determines the relative position of the Fermi level EF . Changing the position of EF turns numerous non-emissive QDs into QDs with more efficient trion emission. γrX− and γnrX− indicate the radiative and non-radiative decay rates of negative trions, while γrX+ and γnrX+ indicate the radiative and non-radiative decay rates of positive trions. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

149

5 Quantum Dot Light-emitting Field-effect Transistors Based on all previous discussions a model for recombination channels in PbS QDs is proposed, explaining the PL intensity and PL average lifetime variations depending on doping conditions (Figure 5.21). The model considers three different charge accumulation regimes, namely initial conditions (no charge accumulation), electron accumulation, and hole accumulation. The exciton emission is represented by the sum of radiative γrX0 and non-radiative γnrX0 decay rates for all QDs. Under initial conditions (middle column, 0 V < Vg < 2 V), there are no mobile charge carriers in the PbS QD thin-film. The non-radiative decay rate is caused by exciton losses via quenching or dissociation, rather than by trap states. As revealed by PL transient analysis, only 1-3 % of the QDs are bright, while the 97-99 % remaining QDs appear dark. The large fraction of dark QDs is caused by dominant non-radiative transitions and fast decay rates (γnr ). The electron accumulation regime is reached by applying a gate voltage above 2 V (left column), and thus shifting the Fermi level towards the CB. As a consequence, electron trap states are deactivated in numerous QDs. This results in QDs with emission from negative trions, composed of radiative (γrX− ) and non-radiative (γnrX− ) decay rates. The assumption is supported by the modified lifetime and the increased intensity of the slow decay component (Figure 5.19 and 5.20). However, the shift of the Fermi level not only induces a transformation of initially dark QDs (“blue”) into bright QDs (“green”) (upper QDs in Figure 5.21), but it also converts some bright QDs (“green”) in dark QDs with an excess of hole traps (“red”) (middle QDs in Figure 5.21), while initially “red” QDs (predominant hole traps) remain dark QDs. A negative gate voltage was applied to enter the hole accumulation regime (last column Figure 5.21, Vg < 0 V), and thus shifting the Fermi level towards the VB. This Fermi level shift towards the VB is the opposite process compared to the Fermi level shift towards the CB and thus induces the same subsequent processes, only vice versa. Light emission occurs through a radiative recombination of positive trions, with radiative γrX+ and non-radiative γnrX+ decay rates. In order to deactivate trap states in PbS QDs, high charge carrier densities must be achieved. These charge carriers cannot only be accumulated by electrical injection but also through an excitation with high laser pumping rates. Figure 5.22 depicts the decay curves for two different excitation power densities per pulse, showing the appearance of a slow decay component with increasing excitation power. This slow decay component is similar to the slow decay component which emerges in consequence of gate biasing. This experiment further supports the notion that increasing the excitation power or applying a gate bias yield the same results. It should be mentioned that electron or hole trap states might also recombine

150

5.3 Charge Carrier Dynamics in PbS Quantum Dot LEFETs radiatively. However, the energy difference for a radiative recombination of midgap states would be smaller than the difference between CB and VB (i.e., smaller than the band gap), which would result in a low-energy emission far beyond the InGaAs detector limit at 1600 nm. For this reason a recombination of midgap states was always considered as non-radiative in the proposed model. PbS 1 IRF 15 -2 0.088x10 cm 15 -2 2.096x10 cm

0.1

0.01 0

10

20

Time (ns)

30

PbS 3 (c)

1

IRF -2 15 0.088x10 cm -2 15 2.096x10 cm

0.1

0.01 0

10

20

Time (ns)

30

Norm. Intensity

(b)

1

Norm. Intensity

Norm. Intensity

(a)

PbS 2 1

IRF 15 -2 0.088x10 cm 15 -2 2.096x10 cm

0.1

0.01 0

10

20

30

Time (ns)

Figure 5.22: Normalized PL decay curves for two different excitation power densities per pulse (“green” = 0.088 · 1015 cm−2 ; “red” = 2.096 · 1015 cm−2 ) for LEFETs fabricated with (a) PbS 1, (b) PbS 2, and (c) PbS 3 QDs. The spectra demonstrate that with an increasing excitation power a slow decay component is emerging. IRF indicates the instrument response function. (Figure adapted from Schornbaum et al.[517] Copyright 2015 American Chemical Society.)

The proposed model states that an increased charge carrier density results in two effects. First, emission is generated by a dominant trion emission and second, many dark QDs with fast non-radiative losses are transformed into bright QDs, thus increasing the overall emission efficiency. The second effect might be primarily responsible for the EQE enhancement with increasing charge carrier densities (Figure 5.13). The introduction of trions in the model is necessary to explain the different emission lifetimes for electron and hole accumulation, which cannot be justified with a simple deactivation of trap states. Since literature reports on trions in QDs are rare and thus a comparison of the proposed trion theory is hardly possible, this model serves as one reasonable explanation for the experimental observations.

5.3.4 Summary In the previous section the charge carrier dynamics in electrolyte-gated PbS QD LEFETs were studied and the effect of high electron and/or hole accumulation on emission efficiencies was investigated. PbS QD thin-films were characterized by gate voltage dependent steady-state and time-resolved PL measurements. These experiments revealed an increase in PL intensity and PL average lifetime with 151

5 Quantum Dot Light-emitting Field-effect Transistors charge accumulation. With the help of non-linear transient analysis it was shown, that the increase in PL average lifetime arises from an emergence of a slow decay component, while fast decay components remain constant. Finally, a model is proposed to explain the effects of charge carrier accumulation on emission efficiencies. Initially, PbS QD thin-film emission is limited by non-radiative decay channels caused by charge carrier trap states. These trap states might be induced during synthesis or the ligand-exchange process. Electrochemical doping is induced by applying a gate voltage, leading to charge carrier accumulation. This charge carrier accumulation results in an increase in PL QY through a deactivation of non-radiative decay channels simultaneously to the creation of trions as radiative decay channels. Thus, the enhancement of PL QY can be explained by a dominant trion emission. Moreover, trion emission also explains the increasing EQE at high current densities, which was demonstrated in Section 5.2.5.

5.4 Conclusion of Chapter 5 It is very important to understand and control charge carrier dynamics at high charge carrier densities in order to improve efficiencies of QD-based optoelectronic devices. Ambipolar LEFETs provide an ideal system to study these processes, however so far LEFETs were only achieved with one-dimensional (e.g., carbon nanotubes), two-dimensional (e.g., MoS2 ), and bulk (e.g., polymers) materials, but no LEFETs with zero-dimensional materials could be realized until now. This chapter presented the first LEFET based on a zero-dimensional material. Electrolyte-gated PbS QD LEFETs, fabricated with ligand-exchanged QD thinfilms exhibit NIR light-emission from a confined region within the channel. The spectral position of the EL is determined by a utilization of different sized PbS QDs for LEFET preparation. Experiments on emission efficiencies revealed a significant enhancement of EQE with current density. Moreover, PL investigations indicated an increase in PL intensity and PL average lifetime with charge carrier accumulation. Based on these experimental results it is deduced, that the initially low emission efficiencies of PbS QD thin-films can be ascribed to non-radiative losses through trap states, while increased emission efficiencies at high charge carrier densities result from a deactivation of trap states and a subsequent dominant trion emission. These findings show that high charge carrier densities can actually improve device performance and therefore are of utmost importance for all QDbased optoelectronic devices, e.g., solar cells and LEDs.

152

6 Conclusion and Outlook This thesis addresses two fundamental challenges that arise when quantum dots (QDs) are used in optoelectronic devices. Chapter 4 dealt with a strategy to improve the overall charge carrier properties in QD-based devices. Therefore, QDs were combined with high charge carrier mobility materials in order to make use of both, the efficient and tunable light absorption properties of QDs and the high charge carrier mobilities of the second material. Chapter 5 concentrated on improving the understanding of the charge carrier dynamics in QD solids. To understand and even more important to subsequently control the charge carrier dynamics, ambipolar light-emitting field-effect transistors (LEFETs) based on PbS QDs were fabricated and charging and emission processes were characterized in detail. Summary of the results. Chapter 4 introduced a new synthesis route, based on an in situ hot-injection synthesis technique, to grow PbSe QDs on various high mobility nanomaterials, achieving a non-covalent coupling. The direct growth ensures an intimate contact between the two materials, leaving no space for linker molecules. Thus, a good electronic coupling between PbSe QDs and a variety of high mobility nanomaterials was realized. The synthesis was first developed and optimized for PbSe-single-walled carbon nanotube (SWNT) hybrids, where PbSe QDs were grown in presence of SWNT bundles. Transmission electron microscopy (TEM) tomography investigations enabled a three-dimensional reconstruction of the hybrid, showing that the QDs use the SWNT bundles as growth template, forming half-ring shaped particles around the bundles. It was found, that although the linkage is non-covalent, the hybrid structure is quite stable and can withstand repeated ultrasonication cycles without a major separation of QDs and SWNT bundles. This remarkable stability is attributed to the stabilization of the PbSe QD dipole moment by the large electronic π-system of the SWNTs. The stabilization of the dipole moment might also be the reason for a favored growth of PbSe QDs on SWNT sidewalls compared to a growth of spherical particles in solution. Highresolution (HR) TEM images revealed, that most of the PbSe QDs were oriented

153

6 Conclusion and Outlook with their {002} lattice planes perpendicular to the SWNT bundles, indicating a preferred orientation. The new developed synthesis route was transferred to a synthesis where SWNTs were replaced with high-mobility layered materials, e.g., graphite/graphene and the transition metal dichalcogenides (TMDs) MoS2 and WS2 . The in situ hot-injection synthesis yielded hybrids of PbSe QDs with few-layer graphene (FLG), and multilayer MoS2 and WS2 flakes, proving the general feasibility of the former developed synthesis technique. It was found that PbSe QDs exhibit a preferred orientation on the MoS2 nanoflakes, with three equivalent orientation variants, verifying an epitaxial growth of PbSe QDs on MoS2 nanoflakes. Cross-sectional HRTEM images clearly showed that PbSe QDs were directly attached to the MoS2 nanoflakes, leaving no space for linker molecules. It was claimed that the oleic acid ligands protect the outer surface of the PbSe QDs, while the interface remains ligand-free, thus enabling a fast charge transfer from PbSe QDs to the MoS2 nanoflake. Probably the hybrid growth is initiated through a Pb2+ nucleation on either SWNT sidewalls or the layered materials and subsequently, right after Se-precursor injection, the QD growth starts. A further growth of the QDs is likely to occur via spontaneous oriented attachment. The nucleation of lead is a crucial step to achieve a dense coverage of the charge transporting material with PbSe QDs. PbSe-MoS2 hybrids were exemplary used to fabricate photodetectors via a simple drop-casting process, both on a rigid and a flexible substrate. The photodetectors showed a photoresponse to near-infrared (NIR) light, verifying an efficient charge separation at the interface between PbSe and MoS2 . The photoresponse is long-time air-stable, even after around fifty on-off switching cycles and several weeks of unprotected storage in air. The flexible devices demonstrated an excellent mechanical stability upon repeated bending and even withstood small bending radii. Chapter 5 focused on the understanding of charge carrier dynamics in PbS QD thin-films at high charge carrier densities in order to control and improve the efficiencies of QD-based optoelectronic devices. The first LEFETs based on a zero-dimensional semiconductor were demonstrated, using ligand-exchanged PbS QD thin-films. High charge carrier densities were achieved through electrolytegating, which made it possible to profit from the ambipolar properties of PbS QDs. Due to this ambipolarity, electrons and holes were simultaneously present in the transistor channel. Electroluminescence (EL) characterization demonstrated NIR light-emission from a confined region within the transistor channel upon recombination of these electrons and holes. LEFETs with different NIR emission wavelengths were achieved by using different sized PbS QDs, demonstrating the

154

spectral tunability through QD size. The determination of external quantum efficiencys (EQEs) revealed an overall low EQE of maximum 0.002 %. However, a significant enhancement with increasing current densities was observed. The accumulation of charge carriers in the QD thin-film induced an increase in photoluminescence (PL) intensity and PL average lifetime. A subsequent non-linear transient analysis revealed that the increase in PL average lifetime arose from the emergence of a slow decay component, while the fast decay components remained constant. Finally, a model was proposed to explain the effects of charge carrier accumulation on emission efficiencies. The initial low emission of PbS QD thin-films was limited by non-radiative losses caused by charge carrier traps. By applying a gate voltage, charge carriers were accumulated and emission was increased. It was stated that PL QY enhancement was caused by a dominant trion emission, which resulted from a deactivation of non-radiative decay channels, and a simultaneously creation of trions as radiative decay channels. Impact of the results and outlook on further experiments. Prior to this work there was no published systematic investigation of the interface in QD hybrids and the influence of a second material on the QD growth mechanism. Within the scope of this thesis, it was found that by choosing the right coupling strategy, a direct, linker-free interface can be achieved. Additionally it was shown that the presence of the second material greatly influences the shape (PbSe-SWNTs) and the orientation of the QDs (PbSe-MoS2 ). Finally it was demonstrated, that photodetectors based on PbSe-MoS2 exhibit a remarkable air-stability, which is uncommon for devices based on PbX QDs (X = S, Se, Te). Further investigations and experiments should involve an optimization of PbSe-SWNTs, in order to be able to use these hybrids in photodetectors. SWNTs exhibit a much higher charge carrier mobility than MoS2 and thus could yield photodetectors with a higher sensitivity. One possibility to achieve these PbSe-SWNT hybrids could be to use only semiconducting SWNTs for hybrid formation and thus avoid high dark currents. Moreover, it would be important to perform a systematic characterization of the PbSe-MoS2 photodetector performance, concentrating on all figures of merit introduced in Chapter 2. As a next step, hybrid materials with two different QDs (e.g., CdS and PbS) could be synthesized in order to extend the spectral detection range into the visible. Before this work, LEFETs were only realized with bulk, two-dimensional, and one-dimensional materials, but the fabrication of LEFETs with a zero-dimensional material was not successful up to this point. Thus, the results of this thesis

155

6 Conclusion and Outlook demonstrate for the first time that the concept of LEFETs is universal and also works for zero-dimensional materials. Moreover, a deep understanding of charge carrier dynamics and charge carrier transport in PbS QD thin-films was provided, including the finding that high charge carrier densities are not always detrimental to device performance, they can even improve it. However, LEFETs still require optimization, e.g. improvement of EQEs, QD thin-film morphology, and device structure. A ligand exchange with short inorganic ligands in contrast to 3-mercaptopropionic acid (MPA) could improve charge carrier mobilities in QD thin-films, leading to higher drain currents and possible higher EQEs. Moreover, LEFETs with oxide dielectrics already revealed promising results and thus a further optimization of this device structure exhibits great potential.

156

Bibliography [1] S. Coe-Sullivan, J. S. Steckel, W.-K. Woo, M. G. Bawendi, V. Bulovi´c, “Large-Area Ordered Quantum-Dot Monolayers via Phase Separation During Spin-Casting”, Adv. Funct. Mater. 2005, 15 (7), 1117–1124. [2] Y. Liu, M. Gibbs, J. Puthussery, S. Gaik, R. Ihly, H. W. Hillhouse, M. Law, “Dependence of Carrier Mobility on Nanocrystal Size and Ligand Length in PbSe Nanocrystal Solids”, Nano Lett. 2010, 10 (5), 1960–1969. [3] J. M. Luther, M. Law, M. C. Beard, Q. Song, M. O. Reese, R. J. Ellingson, A. J. Nozik, “Schottky Solar Cells Based on Colloidal Nanocrystal Films”, Nano Lett. 2008, 8 (10), 3488–3492. [4] G. Konstantatos, I. Howard, A. Fischer, S. Hoogland, J. Clifford, E. Klem, L. Levina, E. H. Sargent, “Ultrasensitive Solution-Cast Quantum Dot Photodetectors”, Nature 2006, 442 (7099), 180–183. [5] J. Tang, G. Konstantatos, S. Hinds, S. Myrskog, A. G. PattantyusAbraham, J. Clifford, E. H. Sargent, “Heavy-Metal-Free Solution-Processed Nanoparticle-Based Photodetectors: Doping of Intrinsic Vacancies Enables Engineering of Sensitivity and Speed”, ACS Nano 2009, 3 (2), 331–338. [6] M. B¨oberl, M. V. Kovalenko, S. Gamerith, E. J. W. List, W. Heiss, “InkjetPrinted Nanocrystal Photodetectors Operating up to 3 µm Wavelengths”, Adv. Mater. 2007, 19 (21), 3574–3578. [7] H. M. Haverinen, R. A. Myllyl¨a, G. E. Jabbour, “Inkjet Printing of Light Emitting Quantum Dots”, Appl. Phys. Lett. 2009, 94 (7), 073108. [8] V. Wood, M. J. Panzer, J. Chen, M. S. Bradley, J. E. Halpert, M. G. Bawendi, V. Bulovi´c, “Inkjet-Printed Quantum Dot-Polymer Composites for Full-Color AC-Driven Displays”, Adv. Mater. 2009, 21 (21), 2151–2155. [9] M. Pehnt, D. L. Schulz, C. J. Curtis, K. M. Jones, D. S. Ginley, “Nanoparticle Precursor Route to Low-Temperature Spray Deposition of CdTe Thin Films”, Appl. Phys. Lett. 1995, 67 (15), 2176–2178. [10] I. J. Kramer, J. C. Minor, G. Moreno-Bautista, L. Rollny, P. Kanjanaboos, D. Kopilovic, S. M. Thon, G. H. Carey, K. W. Chou, D. Zhitomirsky, A. Amassian, E. H. Sargent, “Efficient Spray-Coated Colloidal Quantum Dot Solar Cells”, Adv. Mater. 2015, 27 (1), 116–121. [11] K. Bourzac, “Quantum Dots Go On Display”, Nature 2013, 493 (7432).

157

Bibliography [12] D. A. Wheeler, J. Z. Zhang, “Exciton Dynamics in Semiconductor Nanocrystals”, Adv. Mater. 2013, 25 (21), 2878–2896. [13] M. Reed, J. Randall, R. Aggarwal, R. Matyi, T. Moore, A. Wetsel, “Observation of Discrete Electronic States in a Zero-Dimensional Semiconductor Nanostructure”, Phys. Rev. Lett. 1988, 60 (6), 535–537. [14] F. W. Wise, “Lead Salt Quantum Dots: the Limit of Strong Quantum Confinement”, Acc. Chem. Res. 2000, 33 (11), 773–780. [15] F. Hetsch, N. Zhao, S. V. Kershaw, A. L. Rogach, “Quantum Dot Field Effect Transistors”, Mater. Today 2013, 16 (9), 312–325. [16] Y. Shirasaki, G. J. Supran, M. G. Bawendi, V. Bulovic, “Emergence of Colloidal Quantum-Dot Light-Emitting Technologies”, Nat. Photon. 2013, 7 (1), 13–23. [17] I. Moreels, Y. Justo, B. De Geyter, K. Haustraete, J. C. Martins, Z. Hens, “Size-Tunable, Bright, and Stable PbS Quantum Dots: A Surface Chemistry Study”, ACS Nano 2011, 5 (3), 2004–2012. [18] W. Plugge, C. van der Vliest, “The Use of Near Infrared Spectroscopy in the Quality Control Laboratory of the Pharmaceutical Industry”, J. Pharm. Biomed. Anal. 1992, 10 (10), 797–803. [19] N. Prieto, R. Roehe, P. Lav´ın, G. Batten, S. Andr´es, “Application of Near Infrared Reflectance Spectroscopy to Predict Meat and Meat Products Quality: A Review”, Meat Sci. 2009, 83 (2), 175–186. [20] H. Huang, H. Yu, H. Xu, Y. Ying, “Near Infrared Spectroscopy for On/InLine Monitoring of Quality in Foods and Beverages: A Review”, J. Food Eng. 2008, 87 (3), 303–313. [21] E. H. Sargent, “Infrared Quantum Dots”, Adv. Mater. 2005, 17 (5), 515–522. [22] V. Wood, M. Panzer, S.-C. Sullivan, V. Bulovic, Colloidal Quantum Dot Optoelectronics and Photovoltaics, Cambridge University Press, 2013. [23] Y. T. Lim, S. Kim, A. Nakayama, N. E. Stott, M. G. Bawendi, J. F. Frangioni, “Selection of Quantum Dot Wavelengths for Biomedical Assays and Imaging”, Mol. Imaging 2003, 2 (1), 50–64. [24] F. C. J. M. van Veggel, “Near-Infrared Quantum Dots and Their Delicate Synthesis, Challenging Characterization, and Exciting Potential Applications”, Chem. Mater. 2014, 26 (1), 111–122. [25] A. Rogach, A. Eychmller, S. Hickey, S. Kershaw, “Infrared-Emitting Colloidal Nanocrystals: Synthesis, Assembly, Spectroscopy, and Applications”, Small 2007, 3 (4), 536–557.

158

Bibliography [26] E. H. Sargent, “Solar Cells, Photodetectors, and Optical Sources from Infrared Colloidal Quantum Dots”, Adv. Mater. 2008, 20 (20), 3958–3964. [27] K. A. Abel, J. Shan, J.-C. Boyer, F. Harris, F. C. J. M. van Veggel, “Highly Photoluminescent PbS Nanocrystals: The Beneficial Effect of Trioctylphosphine”, Chem. Mater. 2008, 20 (12), 3794–3796. [28] I. Kang, F. W. Wise, “Electronic Structure and Optical Properties of PbS and PbSe Quantum Dots”, J. Opt. Soc. Am. B 1997, 14 (7), 1632–1646. [29] A. I. Ekimov, A. A. Onushchenko, “Quantum Size Effect in the OpticalSpectra of Semiconductor Microcrystals”, Sov. Phys. Semicond. 1982, 16 (7), 775–778. [30] A. L. Efros, A. L. Efros, “Interband Absorption of Light in a Semiconductor Sphere”, Sov. Phys. Semicond. 1982, 16 (7), 772–775. [31] A. I. Ekimov, A. L. Efros, A. A. Onushchenko, “Quantum Size Effect in Semiconductor Microcrystals”, Solid State Commun. 1985, 56 (11), 921–924. [32] R. Rossetti, S. Nakahara, L. E. Brus, “Quantum Size Effects in the Redox Potentials, Resonance Raman Spectra, and Electronic Spectra of CdS Crystallites in Aqueous Solution”, J. Chem. Phys 1983, 79 (2), 1086–1088. [33] L. E. Brus, “Electron-Electron and Electron-Hole Interactions in Small Semiconductor Crystallites: The Size Dependence of the Lowest Excited Electronic State”, J. Chem. Phys 1984, 80 (9), 4403–4409. [34] L. Brus, “Electronic Wave Functions in Semiconductor Clusters: Experiment and Theory”, J. Phys. Chem. 1986, 90 (12), 2555–2560. [35] H. Weller, U. Koch, M. Guti´errez, A. Henglein, “Photochemistry of Colloidal Metal Sulfides. 7. Absorption and Fluorescence of Extremely Small ZnS Particles (The World of the Neglected Dimensions)”, Ber. Bunsenges. Phys. Chem. 1984, 88 (7), 649–656. [36] A. P. Alivisatos, “Semiconductor Clusters, Nanocrystals, and Quantum Dots”, Science 1996, 271 (5251), 933–937. [37] A. P. Alivisatos, “Perspectives on the Physical Chemistry of Semiconductor Nanocrystals”, J. Phys. Chem. 1996, 100 (31), 13226–13239. [38] T. D. Krauss, J. J. Peterson, Colloidal Quantum Dot Optoelectronics and Photovoltaics, Cambridge University Press, 2013. [39] C. B. Murray, C. R. Kagan, M. G. Bawendi, “Synthesis and Characterization of Monodisperse Nanocrystals and Close-Packed Nanocrystal Assemblies”, Annu. Rev. Mater. Sci. 2000, 30 (1), 545–610.

159

Bibliography [40] B. M. Cossairt, P. Juhas, S. J. L. Billinge, J. S. Owen, “Tuning the Surface Structure and Optical Properties of CdSe Clusters Using Coordination Chemistry”, J. Phys. Chem. Lett. 2011, 2 (24), 3075–3080. [41] H. Weller, H. Schmidt, U. Koch, A. Fojtik, S. Baral, A. Henglein, W. Kunath, K. Weiss, E. Dieman, “Photochemistry of Colloidal Semiconductors. Onset of Light Absorption as a Function of Size of Extremely Small CdS Particles”, Chem. Phys. Lett. 1986, 124 (6), 557–560. [42] L. Spanhel, M. Haase, H. Weller, A. Henglein, “Photochemistry of Colloidal Semiconductors. 20. Surface Modification and Stability of Strong Luminescing CdS Particles”, J. Am. Chem. Soc. 1987, 109 (19), 5649–5655. [43] A. Henglein, “Small-Particle Research: Physicochemical Properties of Extremely Small Colloidal Metal and Semiconductor Particles”, Chem. Rev. 1989, 89 (8), 1861–1873. [44] M. L. Steigerwald, L. E. Brus, “Semiconductor Crystallites: A Class of Large Molecules”, Acc. Chem. Res. 1990, 23 (6), 183–188. [45] A. Eychm¨ uller, “Structure and Photophysics of Semiconductor Nanocrystals”, J. Phys. Chem. B 2000, 104 (28), 6514–6528. [46] C. B. Murray, D. J. Norris, M. G. Bawendi, “Synthesis and Characterization of Nearly Monodisperse CdE (E = Sulfur, Selenium, Tellurium) Semiconductor Nanocrystallites”, J. Am. Chem. Soc. 1993, 115 (19), 8706–8715. [47] W. W. Yu, X. Peng, “Formation of High-Quality CdS and Other II-VI Semiconductor Nanocrystals in Noncoordinating Solvents: Tunable Reactivity of Monomers”, Angew. Chem., Int. Ed. 2002, 41 (13), 2368–2371. [48] Y. Yin, A. P. Alivisatos, “Colloidal Nanocrystal Synthesis and the OrganicInorganic Interface”, Nature 2005, 437 (7059), 664–670. [49] C. de Mello Doneg´a, P. Liljeroth, D. Vanmaekelbergh, “Physicochemical Evaluation of the Hot-Injection Method, a Synthesis Route for Monodisperse Nanocrystals”, Small 2005, 1 (12), 1152–1162. [50] V. K. LaMer, R. H. Dinegar, “Theory, Production and Mechanism of Formation of Monodispersed Hydrosols”, J. Am. Chem. Soc. 1950, 72 (11), 4847–4854. [51] J. Y. Rempel, M. G. Bawendi, K. F. Jensen, “Insights into the Kinetics of Semiconductor Nanocrystal Nucleation and Growth”, J. Am. Chem. Soc. 2009, 131 (12), 4479–4489. [52] L. Qu, W. W. Yu, X. Peng, “In Situ Observation of the Nucleation and Growth of CdSe Nanocrystals”, Nano Lett. 2004, 4 (3), 465–469.

160

Bibliography ¨ [53] W. Ostwald, “Uber die Vermeintliche Isomerie des Roten und Gelben Quecksilberoxyds und die Oberfl¨achenspannung Fester K¨orper”, Z. Phys. Chem. 1900, 34, 495–503. [54] H. Reiss, “The Growth of Uniform Colloidal Dispersions”, J. Chem. Phys 1951, 19 (4), 482–487. [55] X. Peng, J. Wickham, A. P. Alivisatos, “Kinetics of II-VI and III-V Colloidal Semiconductor Nanocrystal Growth: “Focusing” of Size Distributions”, J. Am. Chem. Soc. 1998, 120 (21), 5343–5344. [56] M. I. Bodnarchuk, M. V. Kovalenko, Colloidal Quantum Dot Optoelectronics and Photovoltaics, Cambridge University Press, 2013. [57] A. Hassinen, I. Moreels, K. De Nolf, P. F. Smet, J. C. Martins, Z. Hens, “Short-Chain Alcohols Strip X-Type Ligands and Quench the Luminescence of PbSe and CdSe Quantum Dots, Acetonitrile Does Not”, J. Am. Chem. Soc. 2012, 134 (51), 20705–20712. [58] E. D. Goodwin, B. T. Diroll, S. J. Oh, T. Paik, C. B. Murray, C. R. Kagan, “Effects of Post-Synthesis Processing on CdSe Nanocrystals and Their Solids: Correlation between Surface Chemistry and Optoelectronic Properties”, J. Phys. Chem. C 2014, 118 (46), 27097–27105. [59] Y.-w. Jun, J.-s. Choi, J. Cheon, “Shape Control of Semiconductor and Metal Oxide Nanocrystals through Nonhydrolytic Colloidal Routes”, Angew. Chem., Int. Ed. 2006, 45 (21), 3414–3439. [60] J. Hu, L.-s. Li, W. Yang, L. Manna, L.-w. Wang, A. P. Alivisatos, “Linearly Polarized Emission from Colloidal Semiconductor Quantum Rods”, Science 2001, 292 (5524), 2060–2063. [61] K.-S. Cho, D. V. Talapin, W. Gaschler, C. B. Murray, “Designing PbSe Nanowires and Nanorings through Oriented Attachment of Nanoparticles”, J. Am. Chem. Soc. 2005, 127 (19), 7140–7147. [62] X. Peng, L. Manna, W. Yang, J. Wickham, E. Scher, A. Kadavanich, A. P. Alivisatos, “Shape Control of CdSe Nanocrystals”, Nature 2000, 404 (6773), 59–61. [63] L. Manna, E. C. Scher, A. P. Alivisatos, “Synthesis of Soluble and Processable Rod-, Arrow-, Teardrop-, and Tetrapod-Shaped CdSe Nanocrystals”, J. Am. Chem. Soc. 2000, 122 (51), 12700–12706. [64] A. M. Jawaid, D. J. Asunskis, P. T. Snee, “Shape-Controlled Colloidal Synthesis of Rock-Salt Lead Selenide Nanocrystals”, ACS Nano 2011, 5 (8), 6465–6471.

161

Bibliography [65] D. J. Milliron, S. M. Hughes, Y. Cui, L. Manna, J. Li, L.-W. Wang, A. Paul Alivisatos, “Colloidal Nanocrystal Heterostructures with Linear and Branched Topology”, Nature 2004, 430 (6996), 190–195. [66] D. V. Talapin, J. H. Nelson, E. V. Shevchenko, S. Aloni, B. Sadtler, A. P. Alivisatos, “Seeded Growth of Highly Luminescent CdSe/CdS Nanoheterostructures with Rod and Tetrapod Morphologies”, Nano Lett. 2007, 7 (10), 2951–2959. [67] H. Li, A. G. Kanaras, L. Manna, “Colloidal Branched Semiconductor Nanocrystals: State of the Art and Perspectives”, Acc. Chem. Res. 2013, 46 (7), 1387–1396. [68] R. L. Penn, J. F. Banfield, “Imperfect Oriented Attachment: Dislocation Generation in Defect-Free Nanocrystals”, Science 1998, 281 (5379), 969–971. [69] L. Manna, D. J. Milliron, A. Meisel, E. C. Scher, A. P. Alivisatos, “Controlled Growth of Tetrapod-Branched Inorganic Nanocrystals”, Nat. Mater. 2003, 2 (6), 382–385. [70] M.-S. Kim, Y.-M. Sung, “Successive Solution-Liquid-Solid (SLS) Growth of Heterogeneous Nanowires”, Chem. Mater. 2013, 25 (21), 4156–4164. [71] Y. J. Na, H. S. Kim, J. Park, “Morphology-Controlled Lead Selenide Nanocrystals and Their In Situ Growth on Carbon Nanotubes”, J. Phys. Chem. C 2008, 112 (30), 11218–11226. [72] W. Shi, Y. Sahoo, H. Zeng, Y. Ding, M. T. Swihart, P. N. Prasad, “Anisotropic Growth of PbSe Nanocrystals on Au-Fe3 O4 Hybrid Nanoparticles”, Adv. Mater. 2006, 18 (14), 1889–1894. [73] K.-T. Yong, Y. Sahoo, K. R. Choudhury, M. T. Swihart, J. R. Minter, P. N. Prasad, “Shape Control of PbSe Nanocrystals Using Noble Metal Seed Particles”, Nano Lett. 2006, 6 (4), 709–714. [74] C. Ratanatawanate, C. Xiong, K. J. Balkus, “Fabrication of PbS Quantum Dot Doped TiO2 Nanotubes”, ACS Nano 2008, 2 (8), 1682–1688. [75] R. L. Penn, J. F. Banfield, “Oriented Attachment and Growth, Twinning, Polytypism, and Formation of Metastable Phases: Insights from Nanocrystalline TiO2 ”, Am. Mineral. 1998, 83, 1077–1082. [76] R. L. Penn, J. F. Banfield, “Morphology Development and Crystal Growth in Nanocrystalline Aggregates under Hydrothermal Conditions: Insights from Titania”, Geochim. Cosmochim. Acta 1999, 63 (10), 1549. [77] M. A. van Huis, L. T. Kunneman, K. Overgaag, Q. Xu, G. Pandraud, H. W. Zandbergen, D. Vanmaekelbergh, “Low-Temperature Nanocrystal Unification through Rotations and Relaxations Probed by In Situ Transmission Electron Microscopy”, Nano Lett. 2008, 8 (11), 3959–3963.

162

Bibliography [78] D. Li, M. H. Nielsen, J. R. I. Lee, C. Frandsen, J. F. Banfield, J. J. De Yoreo, “Direction-Specific Interactions Control Crystal Growth by Oriented Attachment”, Science 2012, 336 (6084), 1014–1018. [79] C. Pacholski, A. Kornowski, H. Weller, “Self-Assembly of ZnO: From Nanodots to Nanorods”, Angew. Chem., Int. Ed. 2002, 41 (7), 1188–1191. [80] Z. Tang, N. A. Kotov, M. Giersig, “Spontaneous Organization of Single CdTe Nanoparticles into Luminescent Nanowires”, Science 2002, 297 (5579), 237–240. [81] W. Lu, P. Gao, W. B. Jian, Z. L. Wang, J. Fang, “Perfect Orientation Ordered in-Situ One-Dimensional Self-Assembly of Mn-Doped PbSe Nanocrystals”, J. Am. Chem. Soc. 2004, 126 (45), 14816–14821. [82] M. P. Boneschanscher, W. H. Evers, J. J. Geuchies, T. Altantzis, B. Goris, F. T. Rabouw, S. A. P. van Rossum, H. S. J. van der Zant, L. D. A. Siebbeles, G. Van Tendeloo, I. Swart, J. Hilhorst, A. V. Petukhov, S. Bals, D. Vanmaekelbergh, “Long-Range Orientation and Atomic Attachment of Nanocrystals in 2D Honeycomb Superlattices”, Science 2014, 344 (6190), 1377–1380. [83] Z. A. Peng, X. Peng, “Nearly Monodisperse and Shape-Controlled CdSe Nanocrystals via Alternative Routes: Nucleation and Growth”, J. Am. Chem. Soc. 2002, 124 (13), 3343–3353. [84] X. Peng, “Mechanisms for the Shape-Control and Shape-Evolution of Colloidal Semiconductor Nanocrystals”, Adv. Mater. 2003, 15 (5), 459–463. [85] D. R. Baker, P. V. Kamat, “Tuning the Emission of CdSe Quantum Dots by Controlled Trap Enhancement”, Langmuir 2010, 26 (13), 11272–11276. [86] J. K. Cooper, A. M. Franco, S. Gul, C. Corrado, J. Z. Zhang, “Characterization of Primary Amine Capped CdSe, ZnSe, and ZnS Quantum Dots by FT-IR: Determination of Surface Bonding Interaction and Identification of Selective Desorption”, Langmuir 2011, 27 (13), 8486–8493. [87] A. M. Smith, S. Nie, “Semiconductor Nanocrystals: Structure, Properties, and Band Gap Engineering”, Acc. Chem. Res. 2010, 43 (2), 190–200. [88] G. Kalyuzhny, R. W. Murray, “Ligand Effects on Optical Properties of CdSe Nanocrystals”, J. Phys. Chem. B 2005, 109 (15), 7012–7021. [89] S. Chandramohan, B. D. Ryu, H. K. Kim, C.-H. Hong, E.-K. Suh, “TrapState-Assisted White Light Emission from a CdSe Nanocrystal Integrated Hybrid Light-Emitting Diode”, Opt. Lett. 2011, 36 (6), 802–804. [90] S. Hoogland, Colloidal Quantum Dot Optoelectronics and Photovoltaics, Cambridge University Press, 2013.

163

Bibliography [91] T. Nann, W. M. Skinner, “Quantum Dots for Electro-Optic Devices”, ACS Nano 2011, 5 (7), 5291–5295. [92] G. Konstantatos, E. H. Sargent, “Nanostructured Materials for Photon Detection”, Nat. Nanotechnol. 2010, 5 (6), 391–400. [93] M. A. Schreuder, K. Xiao, I. N. Ivanov, S. M. Weiss, S. J. Rosenthal, “White Light-Emitting Diodes Based on Ultrasmall CdSe Nanocrystal Electroluminescence”, Nano Lett. 2010, 10 (2), 573–576. [94] B. Valeur, M. N. Berberan-Santos, Molecular Fluorescence: Principles and Applications, 2. Aufl., Wiley-VCH Verlag GmbH & Co. KGaA, 2012. [95] O. Chen, J. Zhao, V. P. Chauhan, J. Cui, C. Wong, D. K. Harris, H. Wei, H.-S. Han, D. Fukumura, R. K. Jain, M. G. Bawendi, “Compact High-Quality CdSe-CdS Core-Shell Nanocrystals with Narrow Emission Linewidths and Suppressed Blinking”, Nat. Mater. 2013, 12 (5), 445–451. [96] D. V. Talapin, A. L. Rogach, A. Kornowski, M. Haase, H. Weller, “Highly Luminescent Monodisperse CdSe and CdSe/ZnS Nanocrystals Synthesized in a Hexadecylamine-Trioctylphosphine Oxide-Trioctylphospine Mixture”, Nano Lett. 2001, 1 (4), 207–211. [97] J. Gao, J. C. Johnson, “Charge Trapping in Bright and Dark States of Coupled PbS Quantum Dot Films”, ACS Nano 2012, 6 (4), 3292–3303. [98] S. A. Majetich, A. C. Carter, “Surface Effects on the Optical Properties of Cadmium Selenide Quantum Dots”, J. Phys. Chem. 1993, 97 (34), 8727– 8731. [99] E. Jang, S. Jun, Y. Chung, L. Pu, “Surface Treatment to Enhance the Quantum Efficiency of Semiconductor Nanocrystals”, J. Phys. Chem. B 2004, 108 (15), 4597–4600. [100] H. H.-Y. Wei, C. M. Evans, B. D. Swartz, A. J. Neukirch, J. Young, O. V. Prezhdo, T. D. Krauss, “Colloidal Semiconductor Quantum Dots with Tunable Surface Composition”, Nano Lett. 2012, 12 (9), 4465–4471. [101] J. Jasieniak, P. Mulvaney, “From Cd-Rich to Se-Rich - the Manipulation of CdSe Nanocrystal Surface Stoichiometry”, J. Am. Chem. Soc. 2007, 129 (10), 2841–2848. [102] Y. Yang, W. Rodr´ıguez-C´ordoba, T. Lian, “Multiple Exciton Generation and Dissociation in PbS Quantum Dot-Electron Acceptor Complexes”, Nano Lett. 2012, 12 (8), 4235–4241. [103] B. Liu, C. He, M. Jin, D. Ding, C. Gao, “Time-Resolved Ultrafast Carrier Dynamics in CdTe Quantum Dots Under High Pressure”, Phys. Status Solidi B 2011, 248 (5), 1102–1105.

164

Bibliography [104] R. Schaller, V. Klimov, “High Efficiency Carrier Multiplication in PbSe Nanocrystals: Implications for Solar Energy Conversion”, Phys. Rev. Lett. 2004, 92 (18), 186601. [105] R. J. Ellingson, M. C. Beard, J. C. Johnson, P. Yu, O. I. Micic, A. J. Nozik, A. Shabaev, A. L. Efros, “Highly Efficient Multiple Exciton Generation in Colloidal PbSe and PbS Quantum Dots”, Nano Lett. 2005, 5 (5), 865–871. [106] G. Allan, C. Delerue, “Role of Impact Ionization in Multiple Exciton Generation in PbSe Nanocrystals”, Phys. Rev. B 2006, 73 (20), 205423. [107] A. Shabaev, A. L. Efros, A. J. Nozik, “Multiexciton Generation by a Single Photon in Nanocrystals”, Nano Lett. 2006, 6 (12), 2856–2863. [108] M. C. Beard, A. G. Midgett, M. C. Hanna, J. M. Luther, B. K. Hughes, A. J. Nozik, “Comparing Multiple Exciton Generation in Quantum Dots To Impact Ionization in Bulk Semiconductors: Implications for Enhancement of Solar Energy Conversion”, Nano Lett. 2010, 10 (8), 3019–3027. [109] V. I. Klimov, “Optical Nonlinearities and Ultrafast Carrier Dynamics in Semiconductor Nanocrystals”, J. Phys. Chem. B 2000, 104 (26), 6112–6123. [110] V. I. Klimov, “Mechanisms for Photogeneration and Recombination of Multiexcitons in Semiconductor Nanocrystals: Implications for Lasing and Solar Energy Conversion”, J. Phys. Chem. B 2006, 110 (34), 16827–16845. [111] Y.-S. Park, W. K. Bae, J. M. Pietryga, V. I. Klimov, “Auger Recombination of Biexcitons and Negative and Positive Trions in Individual Quantum Dots”, ACS Nano 2014, 8 (7), 7288–7296. [112] P. P. Jha, P. Guyot-Sionnest, “Trion Decay in Colloidal Quantum Dots”, ACS Nano 2009, 3 (4), 1011–1015. [113] W. Qin, H. Liu, P. Guyot-Sionnest, “Small Bright Charged Colloidal Quantum Dots”, ACS Nano 2014, 8 (1), 283–291. [114] L. A. Padilha, I. Robel, D. C. Lee, P. Nagpal, J. M. Pietryga, V. I. Klimov, “Spectral Dependence of Nanocrystal Photoionization Probability: The Role of Hot-Carrier Transfer”, ACS Nano 2011, 5 (6), 5045–5055. [115] M. Saba, M. Aresti, F. Quochi, M. Marceddu, M. A. Loi, J. Huang, D. V. Talapin, A. Mura, G. Bongiovanni, “Light-Induced Charged and Trap States in Colloidal Nanocrystals Detected by Variable Pulse Rate Photoluminescence Spectroscopy”, ACS Nano 2013, 7 (1), 229–238. [116] M. Law, J. M. Luther, Q. Song, B. K. Hughes, C. L. Perkins, A. J. Nozik, “Structural, Optical, and Electrical Properties of PbSe Nanocrystal Solids Treated Thermally or with Simple Amines”, J. Am. Chem. Soc. 2008, 130 (18), 5974–5985.

165

Bibliography [117] S. J. Oh, N. E. Berry, J.-H. Choi, E. A. Gaulding, H. Lin, T. Paik, B. T. Diroll, S. Muramoto, C. B. Murray, C. R. Kagan, “Designing High-Performance PbS and PbSe Nanocrystal Electronic Devices through Stepwise, Post-Synthesis, Colloidal Atomic Layer Deposition”, Nano Lett. 2014, 14 (3), 1559–1566. [118] J.-H. Choi, A. T. Fafarman, S. J. Oh, D.-K. Ko, D. K. Kim, B. T. Diroll, S. Muramoto, J. G. Gillen, C. B. Murray, C. R. Kagan, “Bandlike Transport in Strongly Coupled and Doped Quantum Dot Solids: A Route to HighPerformance Thin-Film Electronics”, Nano Lett. 2012, 12 (5), 2631–2638. [119] D. V. Talapin, C. B. Murray, “PbSe Nanocrystal Solids for n- and p-Channel Thin Film Field-Effect Transistors”, Science 2005, 310 (5745), 86–89. [120] E. Talgorn, Y. Gao, M. Aerts, L. T. Kunneman, J. M. Schins, S. T. J., M. A. van Huis, H. S. J. van der Zant, A. J. Houtepen, L. D. A. Siebbeles, “Unity Quantum Yield of Photogenerated Charges and Band-Like Transport in Quantum-Dot Solids”, Nat. Nanotechnol. 2011, 6 (11), 733–739. [121] P. Guyot-Sionnest, C. Wang, “Fast Voltammetric and Electrochromic Response of Semiconductor Nanocrystal Thin Films”, J. Phys. Chem. B 2003, 107 (30), 7355–7359. [122] M. Jarosz, V. Porter, B. Fisher, M. Kastner, M. Bawendi, “Photoconductivity Studies of Treated CdSe Quantum Dot Films Exhibiting Increased Exciton Ionization Efficiency”, Phys. Rev. B 2004, 70 (19), 195327. [123] J. M. Luther, M. Law, Q. Song, C. L. Perkins, M. C. Beard, A. J. Nozik, “Structural, Optical, and Electrical Properties of Self-Assembled Films of PbSe Nanocrystals Treated with 1,2-Ethanedithiol”, ACS Nano 2008, 2 (2), 271–280. [124] E. Talgorn, E. Moysidou, R. D. Abellon, T. J. Savenije, A. Goossens, A. J. Houtepen, L. D. A. Siebbeles, “Highly Photoconductive CdSe Quantum-Dot Films: Influence of Capping Molecules and Film Preparation Procedure”, J. Phys. Chem. C 2010, 114 (8), 3441–3447. [125] Y. Gao, M. Aerts, C. S. S. Sandeep, E. Talgorn, T. J. Savenije, S. Kinge, L. D. A. Siebbeles, A. J. Houtepen, “Photoconductivity of PbSe QuantumDot Solids: Dependence on Ligand Anchor Group and Length”, ACS Nano 2012, 6 (11), 9606–9614. [126] K. S. Jeong, J. Tang, H. Liu, J. Kim, A. W. Schaefer, K. Kemp, L. Levina, X. Wang, S. Hoogland, R. Debnath, L. Brzozowski, E. H. Sargent, J. B. Asbury, “Enhanced Mobility-Lifetime Products in PbS Colloidal Quantum Dot Photovoltaics”, ACS Nano 2012, 6 (1), 89–99. [127] J. M. Luther, J. Gao, M. T. Lloyd, O. E. Semonin, M. C. Beard, A. J. Nozik, “Stability Assessment on a 3% Bilayer PbS/ZnO Quantum Dot Heterojunction Solar Cell”, Adv. Mater. 2010, 22 (33), 3704–3707.

166

Bibliography [128] J. Tang, L. Brzozowski, D. A. R. Barkhouse, X. Wang, R. Debnath, R. Wolowiec, E. Palmiano, L. Levina, A. G. Pattantyus-Abraham, D. Jamakosmanovic, E. H. Sargent, “Quantum Dot Photovoltaics in the Extreme Quantum Confinement Regime: The Surface-Chemical Origins of Exceptional Air- and Light-Stability”, ACS Nano 2010, 4 (2), 869–878. [129] G. I. Koleilat, L. Levina, H. Shukla, S. H. Myrskog, S. Hinds, A. G. Pattantyus-Abraham, E. H. Sargent, “Efficient, Stable Infrared Photovoltaics Based on Solution-Cast Colloidal Quantum Dots”, ACS Nano 2008, 2 (5), 833–840. [130] J. Tang, K. W. Kemp, S. Hoogland, K. S. Jeong, H. Liu, L. Levina, M. Furukawa, X. Wang, R. Debnath, D. Cha, K. W. Chou, A. Fischer, A. Amassian, J. B. Asbury, E. H. Sargent, “Colloidal-Quantum-Dot Photovoltaics Using Atomic-Ligand Passivation”, Nat. Mater. 2011, 10 (10), 765–771. [131] E. J. D. Klem, H. Shukla, S. Hinds, D. D. MacNeil, L. Levina, E. H. Sargent, “Impact of Dithiol Treatment and Air Annealing on the Conductivity, Mobility, and Hole Density in PbS Colloidal Quantum Dot Solids”, Appl. Phys. Lett. 2008, 92 (21), 212105. [132] K. S. Leschkies, T. J. Beatty, M. S. Kang, D. J. Norris, E. S. Aydil, “Solar Cells Based on Junctions between Colloidal PbSe Nanocrystals and Thin ZnO Films”, ACS Nano 2009, 3 (11), 3638–3648. [133] M. S. Kang, A. Sahu, D. J. Norris, C. D. Frisbie, “Size- and TemperatureDependent Charge Transport in PbSe Nanocrystal Thin Films”, Nano Lett. 2011, 11 (9), 3887–3892. [134] P. Nagpal, V. I. Klimov, “Role of Mid-Gap States in Charge Transport and Photoconductivity in Semiconductor Nanocrystal Films”, Nat. Commun. 2011, 2, 486. [135] T. P. Osedach, N. Zhao, T. L. Andrew, P. R. Brown, D. D. Wanger, D. B. Strasfeld, L.-Y. Chang, M. G. Bawendi, V. Bulovi´c, “Bias-Stress Effect in 1,2-Ethanedithiol-Treated PbS Quantum Dot Field-Effect Transistors”, ACS Nano 2012, 6 (4), 3121–3127. [136] D. Yu, C. Wang, P. Guyot-Sionnest, “n-Type Conducting CdSe Nanocrystal Solids”, Science 2003, 300 (5623), 1277–1280. [137] S. Geyer, V. Porter, J. Halpert, T. Mentzel, M. Kastner, M. Bawendi, “Charge Transport in Mixed CdSe and CdTe Colloidal Nanocrystal Films”, Phys. Rev. B 2010, 82 (15), 155201. [138] D. C. Oertel, M. G. Bawendi, A. C. Arango, V. Bulovi´c, “Photodetectors Based on Treated CdSe Quantum-Dot Films”, Appl. Phys. Lett. 2005, 87 (21), 213505.

167

Bibliography [139] B. Sun, A. T. Findikoglu, M. Sykora, D. J. Werder, V. I. Klimov, “Hybrid Photovoltaics Based on Semiconductor Nanocrystals and Amorphous Silicon”, Nano Lett. 2009, 9 (3), 1235–1241. [140] S. K. Poznyak, N. P. Osipovich, A. Shavel, D. V. Talapin, M. Gao, A. Eychm¨ uller, N. Gaponik, “Size-Dependent Electrochemical Behavior of ThiolCapped CdTe Nanocrystals in Aqueous Solution”, J. Phys. Chem. B 2005, 109 (3), 1094–1100. [141] A. G. Pattantyus-Abraham, I. J. Kramer, A. R. Barkhouse, X. Wang, G. Konstantatos, R. Debnath, L. Levina, I. Raabe, M. K. Nazeeruddin, M. Gr¨atzel, E. H. Sargent, “Depleted-Heterojunction Colloidal Quantum Dot Solar Cells”, ACS Nano 2010, 4 (6), 3374–3380. [142] M. V. Kovalenko, M. Scheele, D. V. Talapin, “Colloidal Nanocrystals with Molecular Metal Chalcogenide Surface Ligands”, Science 2009, 324 (5933), 1417–1420. [143] M. V. Kovalenko, B. Spokoyny, J.-S. Lee, M. Scheele, A. Weber, S. Perera, D. Landry, D. V. Talapin, “Semiconductor Nanocrystals Functionalized with Antimony Telluride Zintl Ions for Nanostructured Thermoelectrics”, J. Am. Chem. Soc. 2010, 132 (19), 6686–6695. [144] M. V. Kovalenko, M. I. Bodnarchuk, J. Zaumseil, J.-S. Lee, D. V. Talapin, “Expanding the Chemical Versatility of Colloidal Nanocrystals Capped with Molecular Metal Chalcogenide Ligands”, J. Am. Chem. Soc. 2010, 132 (29), 10085–10092. [145] J.-S. Lee, M. V. Kovalenko, J. Huang, D. S. Chung, D. V. Talapin, “BandLike Transport, High Electron Mobility and High Photoconductivity in All-Inorganic Nanocrystal Arrays”, Nat. Nanotechnol. 2011, 6 (6), 348–352. [146] W. Liu, J.-S. Lee, D. V. Talapin, “III-V Nanocrystals Capped with Molecular Metal Chalcogenide Ligands: High Electron Mobility and Ambipolar Photoresponse”, J. Am. Chem. Soc. 2013, 135 (4), 1349–1357. [147] A. Llordes, A. T. Hammack, R. Buonsanti, R. Tangirala, S. Aloni, B. A. Helms, D. J. Milliron, “Polyoxometalates and Colloidal Nanocrystals as Building Blocks for Metal Oxide Nanocomposite Films”, J. Mater. Chem. 2011, 21 (31), 11631–11638. [148] J. Huang, W. Liu, D. S. Dolzhnikov, L. Protesescu, M. V. Kovalenko, B. Koo, S. Chattopadhyay, E. V. Shenchenko, D. V. Talapin, “Surface Functionalization of Semiconductor and Oxide Nanocrystals with Small Inorganic Oxoanions (PO4 3 – , MoO4 2 – ) and Polyoxometalate Ligands”, ACS Nano 2014, 8 (9), 9388–9402. [149] D. N. Dirin, S. Dreyfuss, M. I. Bodnarchuk, G. Nedelcu, P. Papagiorgis, G. Itskos, M. V. Kovalenko, “Lead Halide Perovskites and Other Metal

168

Bibliography Halide Complexes As Inorganic Capping Ligands for Colloidal Nanocrystals”, J. Am. Chem. Soc. 2014, 136 (18), 6550–6553. [150] H. Zhang, J. Jang, W. Liu, D. V. Talapin, “Colloidal Nanocrystals with Inorganic Halide, Pseudohalide, and Halometallate Ligands”, ACS Nano 2014, 8 (7), 7359–7369. [151] A. T. Fafarman, W.-k. Koh, B. T. Diroll, D. K. Kim, D.-K. Ko, S. J. Oh, X. Ye, V. Doan-Nguyen, M. R. Crump, D. C. Reifsnyder, C. B. Murray, C. R. Kagan, “Thiocyanate-Capped Nanocrystal Colloids: Vibrational Reporter of Surface Chemistry and Solution-Based Route to Enhanced Coupling in Nanocrystal Solids”, J. Am. Chem. Soc. 2011, 133 (39), 15753–15761. [152] D. K. Kim, Y. Lai, B. T. Diroll, C. B. Murray, C. R. Kagan, “Flexible and Low-Voltage Integrated Circuits Constructed from High-Performance Nanocrystal Transistors”, Nat. Commun. 2012, 3, 1216. [153] W.-k. Koh, S. R. Saudari, A. T. Fafarman, C. R. Kagan, C. B. Murray, “Thiocyanate-Capped PbS Nanocubes: Ambipolar Transport Enables Quantum Dot Based Circuits on a Flexible Substrate”, Nano Lett. 2011, 11 (11), 4764–4767. [154] A. Nag, M. V. Kovalenko, J.-S. Lee, W. Liu, B. Spokoyny, D. V. Talapin, “Metal-Free Inorganic Ligands for Colloidal Nanocrystals: S2 – , HS – , Se2 – , HSe – , Te2 – , HTe – , TeS3 2 – , OH – , and NH2 – as Surface Ligands”, J. Am. Chem. Soc. 2011, 133 (27), 10612–10620. [155] D. S. Chung, J.-S. Lee, J. Huang, A. Nag, S. Ithurria, D. V. Talapin, “Low Voltage, Hysteresis Free, and High Mobility Transistors from All-Inorganic Colloidal Nanocrystals”, Nano Lett. 2012, 12 (4), 1813–1820. [156] Z. Ning, Y. Ren, S. Hoogland, O. Voznyy, L. Levina, P. Stadler, X. Lan, D. Zhitomirsky, E. H. Sargent, “All-Inorganic Colloidal Quantum Dot Photovoltaics Employing Solution-Phase Halide Passivation”, Adv. Mater. 2012, 24 (47), 6295–6299. [157] W. K. Bae, J. Joo, L. A. Padilha, J. Won, D. C. Lee, Q. Lin, W.-k. Koh, H. Luo, V. I. Klimov, J. M. Pietryga, “Highly Effective Surface Passivation of PbSe Quantum Dots through Reaction with Molecular Chlorine”, J. Am. Chem. Soc. 2012, 134 (49), 20160–20168. [158] B. A. Ridley, B. Nivi, J. M. Jacobson, “All-Inorganic Field Effect Transistors Fabricated by Printing”, Science 1999, 286 (5440), 746–749. [159] M. Drndi´c, M. V. Jarosz, N. Y. Morgan, M. A. Kastner, M. G. Bawendi, “Transport Properties of Annealed CdSe Colloidal Nanocrystal Solids”, J. Appl. Phys. 2002, 92 (12), 7498–7503.

169

Bibliography [160] L. Turyanska, U. Elfurawi, M. Li, M. W. Fay, N. R. Thomas, S. Mann, J. H. Blokland, P. C. M. Christianen, A. Patan`e, “Tailoring the Physical Properties of Thiol-Capped PbS Quantum Dots by Thermal Annealing”, Nanotechnology 2009, 20 (31), 315604. [161] S. J. Baik, K. Kim, K. S. Lim, S. Jung, Y.-C. Park, D. G. Han, S. Lim, S. Yoo, S. Jeong, “Low-Temperature Annealing for Highly Conductive Lead Chalcogenide Quantum Dot Solids”, J. Phys. Chem. C 2011, 115 (3), 607– 612. [162] A. Sahu, M. S. Kang, A. Kompch, C. Notthoff, A. W. Wills, D. Deng, M. Winterer, C. D. Frisbie, D. J. Norris, “Electronic Impurity Doping in CdSe Nanocrystals”, Nano Lett. 2012, 12 (5), 2587–2594. [163] A. W. Wills, M. S. Kang, K. M. Wentz, S. E. Hayes, A. Sahu, W. L. Gladfelter, D. J. Norris, “Synthesis and Characterization of Al- and In-Doped CdSe Nanocrystals”, J. Mater. Chem. 2012, 22 (13), 6335–6342. [164] M. S. Kang, A. Sahu, C. D. Frisbie, D. J. Norris, “Influence of Silver Doping on Electron Transport in Thin Films of PbSe Nanocrystals”, Adv. Mater. 2013, 25 (5), 725–731. [165] W. J. Baumgardner, K. Whitham, T. Hanrath, “Confined-but-Connected Quantum Solids via Controlled Ligand Displacement”, Nano Lett. 2013, 13 (7), 3225–3231. [166] C. S. S. Sandeep, J. M. Azpiroz, W. H. Evers, S. C. Boehme, I. Moreels, S. Kinge, L. D. A. Siebbeles, I. Infante, A. J. Houtepen, “Epitaxially Connected PbSe Quantum-Dot Films: Controlled Neck Formation and Optoelectronic Properties”, ACS Nano 2014, 8 (11), 11499–11511. [167] A. Roest, J. Kelly, D. Vanmaekelbergh, E. Meulenkamp, “Staircase in the Electron Mobility of a ZnO Quantum Dot Assembly due to Shell Filling”, Phys. Rev. Lett. 2002, 89 (3), 036801. [168] A. J. Houtepen, D. Vanmaekelbergh, “Orbital Occupation in ElectronCharged CdSe Quantum-Dot Solids”, J. Phys. Chem. B 2005, 109 (42), 19634–19642. [169] R. Chandler, A. Houtepen, J. Nelson, D. Vanmaekelbergh, “Electron Transport in Quantum Dot Solids: Monte Carlo Simulations of the Effects of Shell Filling, Coulomb Repulsions, and Site Disorder”, Phys. Rev. B 2007, 75 (8), 085325. [170] A. J. Houtepen, D. Kockmann, D. Vanmaekelbergh, “Reappraisal of VariableRange Hopping in Quantum-Dot Solids”, Nano Lett. 2008, 8 (10), 3516–3520. [171] B. L. Wehrenberg, D. Yu, J. Ma, P. Guyot-Sionnest, “Conduction in Charged PbSe Nanocrystal Films”, J. Phys. Chem. B 2005, 109 (43), 20192–20199.

170

Bibliography [172] D. Yu, C. Wang, B. Wehrenberg, P. Guyot-Sionnest, “Variable Range Hopping Conduction in Semiconductor Nanocrystal Solids”, Phys. Rev. Lett. 2004, 92 (21), 216802. [173] W. W. Yu, J. C. Falkner, B. S. Shih, V. L. Colvin, “Preparation and Characterization of Monodisperse PbSe Semiconductor Nanocrystals in a Noncoordinating Solvent”, Chem. Mater. 2004, 16 (17), 3318–3322. [174] T. Mentzel, V. Porter, S. Geyer, K. MacLean, M. Bawendi, M. Kastner, “Charge Transport in PbSe Nanocrystal Arrays”, Phys. Rev. B 2008, 77 (7), 075316. [175] S. J. Oh, Z. Wang, N. E. Berry, J.-H. Choi, T. Zhao, E. A. Gaulding, T. Paik, Y. Lai, C. B. Murray, C. R. Kagan, “Engineering Charge Injection and Charge Transport for High Performance PbSe Nanocrystal Thin Film Devices and Circuits”, Nano Lett. 2014, 14 (11), 6210–6216. [176] A. Shabaev, A. L. Efros, A. L. Efros, “Dark and Photo-Conductivity in Ordered Array of Nanocrystals”, Nano Lett. 2013, 13 (11), 5454–5461. [177] J. Jang, W. Liu, J. S. Son, D. V. Talapin, “Temperature-Dependent Hall and Field-Effect Mobility in Strongly Coupled All-Inorganic Nanocrystal Arrays”, Nano Lett. 2014, 14 (2), 653–662. [178] R. Ihly, J. Tolentino, Y. Liu, M. Gibbs, M. Law, “The Photothermal Stability of PbS Quantum Dot Solids”, ACS Nano 2011, 5 (10), 8175–8186. [179] N. F. Mott, “Conduction in Non-Crystalline Materials”, Philos. Mag. 1969, 19 (160), 835–852. [180] T. Sakanoue, H. Sirringhaus, “Band-Like Temperature Dependence of Mobility in a Solution-Processed Organic Semiconductor”, Nat. Mater. 2010, 9 (9), 736–740. [181] S. Wang, M. Ha, M. Manno, C. Daniel Frisbie, C. Leighton, “Hopping Transport and the Hall Effect Nar the Insulator-Metal Transition in Electrochemically Gated Poly(3-Hexylthiophene) Transistors”, Nat. Commun. 2012, 3, 1210. [182] C. B. Murray, S. Sun, W. Gaschler, H. Doyle, T. A. Betley, C. R. Kagan, “Colloidal Synthesis of Nanocrystals and Nanocrystal Superlattices”, IBM J. Res. Dev. 2001, 45 (1), 47–56. [183] M. Hines, G. Scholes, “Colloidal PbS Nanocrystals with Size-Tunable NearInfrared Emission: Observation of Post-Synthesis Self-Narrowing of the Particle Size Distribution”, Adv. Mater. 2003, 15 (21), 1844–1849. [184] L. Cademartiri, J. Bertolotti, R. Sapienza, D. S. Wiersma, G. von Freymann, G. A. Ozin, “Multigram Scale, Solventless, and Diffusion-Controlled Route to Highly Monodisperse PbS Nanocrystals”, J. Phys. Chem. B. 2006, 110 (2), 671–673.

171

Bibliography [185] B. L. Wehrenberg, C. Wang, P. Guyot-Sionnest, “Interband and Intraband Optical Studies of PbSe Colloidal Quantum Dots”, J. Phys. Chem. B 2002, 106 (41), 10634–10640. [186] H. Du, C. Chen, R. Krishnan, T. D. Krauss, J. M. Harbold, F. W. Wise, M. G. Thomas, J. Silcox, “Optical Properties of Colloidal PbSe Nanocrystals”, Nano Lett. 2002, 2 (11), 1321–1324. [187] W. Lu, J. Fang, K. L. Stokes, J. Lin, “Shape Evolution and Self Assembly of Monodisperse PbTe Nanocrystals”, J. Am. Chem. Soc. 2004, 126 (38), 11798–11799. [188] J. J. Urban, D. V. Talapin, E. V. Shevchenko, C. B. Murray, “Self-Assembly of PbTe Quantum Dots into Nanocrystal Superlattices and Glassy Films”, J. Am. Chem. Soc. 2006, 128 (10), 3248–3255. [189] J. E. Murphy, M. C. Beard, A. G. Norman, S. P. Ahrenkiel, J. C. Johnson, P. Yu, O. I. Mi´ci´c, R. J. Ellingson, A. J. Nozik, “PbTe Colloidal Nanocrystals: Synthesis, Characterization, and Multiple Exciton Generation”, J. Am. Chem. Soc. 2006, 128 (10), 3241–3247. [190] J. Machol, F. Wise, R. Patel, D. Tanner, “Vibronic Quantum Beats in PbS Microcrystallites”, Phys. Rev. B 1993, 48 (4), 2819–2822. [191] I. Moreels, K. Lambert, D. Smeets, D. De Muynck, T. Nollet, J. C. Martins, F. Vanhaecke, A. Vantomme, C. Delerue, G. Allan, Z. Hens, “Size-Dependent Optical Properties of Colloidal PbS Quantum Dots”, ACS Nano 2009, 3 (10), 3023–3030. [192] I. Moreels, B. Fritzinger, J. C. Martins, Z. Hens, “Surface Chemistry of Colloidal PbSe Nanocrystals”, J. Am. Chem. Soc. 2008, 130 (45), 15081– 15086. [193] M. T. Trinh, A. J. Houtepen, J. M. Schins, J. Piris, L. D. A. Siebbeles, “Nature of the Second Optical Transition in PbSe Nanocrystals”, Nano Lett. 2008, 8 (7), 2112–2117. [194] O. E. Semonin, J. C. Johnson, J. M. Luther, A. G. Midgett, A. J. Nozik, M. C. Beard, “Absolute Photoluminescence Quantum Yields of IR-26 Dye, PbS, and PbSe Quantum Dots”, J. Phys. Chem. Lett. 2010, 1 (16), 2445–2450. [195] C. Fang, M. A. van Huis, D. Vanmaekelbergh, H. W. Zandbergen, “Energetics of Polar and Nonpolar Facets of PbSe Nanocrystals from Theory and Experiment”, ACS Nano 2010, 4 (1), 211–218. [196] M. A. Boles, D. V. Talapin, “Connecting the Dots”, Science 2014, 344 (6190), 1340–1341.

172

Bibliography [197] I. Moreels, K. Lambert, D. De Muynck, F. Vanhaecke, D. Poelman, J. C. Martins, G. Allan, Z. Hens, “Composition and Size-Dependent Extinction Coefficient of Colloidal PbSe Quantum Dots”, Chem. Mater. 2007, 19 (25), 6101–6106. [198] Q. Dai, Y. Wang, X. Li, Y. Zhang, D. J. Pellegrino, M. Zhao, B. Zou, J. Seo, Y. Wang, W. W. Yu, “Size-Dependent Composition and Molar Extinction Coefficient of PbSe Semiconductor Nanocrystals”, ACS Nano 2009, 3 (6), 1518–1524. [199] D. I. Garcia-Gutierrez, D. F. Garcia-Gutierrez, L. M. De Leon-Covian, M. T. Trevi˜ no Gonzalez, M. A. Garza-Navarro, I. E. Moreno-Cortez, R. F. Cienfuegos-Pelaes, “Aberration Corrected STEM Study of the Surface of Lead Chalcogenide Nanoparticles”, J. Phys. Chem. C 2014, 118 (38), 22291– 22298. [200] D. Kim, D.-H. Kim, J.-H. Lee, J. Grossman, “Impact of Stoichiometry on the Electronic Structure of PbS Quantum Dots”, Phys. Rev. Lett. 2013, 110 (19), 196802. [201] K. S. Leschkies, M. S. Kang, E. S. Aydil, D. J. Norris, “Influence of Atmospheric Gases on the Electrical Properties of PbSe Quantum-Dot Films”, J. Phys. Chem. C 2010, 114 (21), 9988–9996. [202] M. Sykora, A. Y. Koposov, J. A. McGuire, R. K. Schulze, O. Tretiak, J. M. Pietryga, V. I. Klimov, “Effect of Air Exposure on Surface Properties, Electronic Structure, and Carrier Relaxation in PbSe Nanocrystals”, ACS Nano 2010, 4 (4), 2021–2034. [203] E. E. Foos, “The Complex Interaction of Spectroscopic Shifts and Electronic Properties in Semiconductor Nanocrystal Films”, J. Phys. Chem. Lett. 2013, 4 (4), 625–632. [204] A. H. Ip, S. M. Thon, S. Hoogland, O. Voznyy, D. Zhitomirsky, R. Debnath, L. Levina, L. R. Rollny, G. H. Carey, A. Fischer, K. W. Kemp, I. J. Kramer, Z. Ning, A. J. Labelle, K. W. Chou, A. Amassian, E. H. Sargent, “Hybrid Passivated Colloidal Quantum Dot Solids”, Nat. Nanotechnol. 2012, 7 (9), 577–582. [205] Z. Ning, O. Voznyy, J. Pan, S. Hoogland, V. Adinolfi, J. Xu, M. Li, A. R. Kirmani, J.-P. Sun, J. Minor, K. W. Kemp, H. Dong, L. Rollny, A. Labelle, G. Carey, B. Sutherland, I. Hill, A. Amassian, H. Liu, J. Tang, O. M. Bakr, E. H. Sargent, “Air-Stable n-Type Colloidal Quantum Dot Solids”, Nat. Mater. 2014, 13 (8), 822–828. [206] M. C. Weidman, M. E. Beck, R. S. Hoffman, F. Prins, W. A. Tisdale, “Monodisperse, Air-Stable PbS Nanocrystals via Precursor Stoichiometry Control”, ACS Nano 2014, 8 (6), 6363–6371.

173

Bibliography [207] M. Yuan, K. W. Kemp, S. M. Thon, J. Y. Kim, K. W. Chou, A. Amassian, E. H. Sargent, “High-Performance Quantum-Dot Solids via Elemental Sulfur Synthesis”, Adv. Mater. 2014, 26 (21), 3513–3519. [208] W. Lu, J. Fang, Y. Ding, Z. L. Wang, “Formation of PbSe Nanocrystals: A Growth toward Nanocubes”, J. Phys. Chem. B 2005, 109 (41), 19219–19222. [209] J. H. Warner, H. Cao, “Shape Control of PbS Nanocrystals Using Multiple Surfactants”, Nanotechnology 2008, 19 (30), 305605. [210] P. Schapotschnikow, M. A. van Huis, H. W. Zandbergen, D. Vanmaekelbergh, T. J. H. Vlugt, “Morphological Transformations and Fusion of PbSe Nanocrystals Studied Using Atomistic Simulations”, Nano Lett. 2010, 10 (10), 3966–3971. [211] B. K. Hughes, J. L. Blackburn, D. Kroupa, A. Shabaev, S. C. Erwin, A. L. Efros, A. J. Nozik, J. M. Luther, M. C. Beard, “Synthesis and Spectroscopy of PbSe Fused Quantum-Dot Dimers”, J. Am. Chem. Soc. 2014, 136 (12), 4670–4679. [212] S. Iijima, “Helical Microtubules of Graphitic Carbon”, Nature 1991, 354 (6348), 56–58. [213] L. Vaisman, H. D. Wagner, G. Marom, “The Role of Surfactants in Dispersion of Carbon Nanotubes”, Adv. Colloid Interface Sci. 2006, 128-130 (0), 37–46. [214] P. Avouris, Z. Chen, V. Perebeinos, “Carbon-Based Electronics”, Nat. Nanotechnol. 2007, 2 (10), 605–615. [215] R. H. Baughman, A. A. Zakhidov, W. A. de Heer, “Carbon Nanotubes–the Route Toward Applications”, Science 2002, 297 (5582), 787–792. [216] S. Iijima, T. Ichihashi, “Single-Shell Carbon Nanotubes of 1-nm Diameter”, Nature 1993, 363 (6430), 603–605. [217] R. Saito, M. Fujita, G. Dresselhaus, M. S. Dresselhaus, “Electronic Structure of Chiral Graphene Tubules”, Appl. Phys. Lett. 1992, 60 (18), 2204–2206. [218] N. Hamada, S.-i. Sawada, A. Oshiyama, “New One-Dimensional Conductors: Graphitic Microtubules”, Phys. Rev. Lett. 1992, 68 (10), 1579–1581. [219] T. W. Odom, J.-L. Huang, P. Kim, C. M. Lieber, “Atomic Structure and Electronic Properties of Single-Walled Carbon Nanotubes”, Nature 1998, 391 (6662), 62–64. [220] J. W. G. Wilder, L. C. Venema, A. G. Rinzler, R. E. Smalley, C. Dekker, “Electronic Structure of Atomically Resolved Carbon Nanotubes”, Nature 1998, 391 (6662), 59–62. [221] S. Frank, P. Poncharal, Z. L. Wang, W. A. d. Heer, “Carbon Nanotube Quantum Resistors”, Science 1998, 280 (5370), 1744–1746.

174

Bibliography [222] C. Biswas, Y. H. Lee, “Graphene Versus Carbon Nanotubes in Electronic Devices”, Adv. Funct. Mater. 2011, 21 (20), 3806–3826. [223] F. Jakubka, Dissertation, University of Erlangen-N¨ urnberg, 2015. [224] D. S. Bethune, C. H. Klang, M. S. de Vries, G. Gorman, R. Savoy, J. Vazquez, R. Beyers, “Cobalt-Catalysed Growth of Carbon Nanotubes with SingleAtomic-Layer Walls”, Nature 1993, 363 (6430), 605–607. [225] A. Thess, R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C. Xu, Y. H. Lee, S. G. Kim, A. G. Rinzler, D. T. Colbert, G. E. Scuseria, D. Tom´anek, J. E. Fischer, R. E. Smalley, “Crystalline Ropes of Metallic Carbon Nanotubes”, Science 1996, 273 (5274), 483–487. [226] M. Jos´e-Yacam´an, M. Miki-Yoshida, L. Rend´on, J. G. Santiesteban, “Catalytic Growth of Carbon Microtubules with Fullerene Structure”, Appl. Phys. Lett. 1993, 62 (6), 657–659. [227] A. M. Cassell, J. A. Raymakers, J. Kong, H. Dai, “Large Scale CVD Synthesis of Single-Walled Carbon Nanotubes”, J. Phys. Chem. B 1999, 103 (31), 6484– 6492. [228] J. H. Hafner, M. J. Bronikowski, B. R. Azamian, P. Nikolaev, A. G. Rinzler, D. T. Colbert, K. A. Smith, R. E. Smalley, “Catalytic Growth of Single-Wall Carbon Nanotubes From Metal Particles”, Chem. Phys. Lett. 1998, 296 (1-2), 195–202. [229] D.-M. Sun, C. Liu, W.-C. Ren, H.-M. Cheng, “A Review of Carbon Nanotubeand Graphene-Based Flexible Thin-Film Transistors”, Small 2013, 9 (8), 1188–1205. [230] M. J. Bronikowski, P. A. Willis, D. T. Colbert, K. A. Smith, R. E. Smalley, “Gas-Phase Production of Carbon Single-Walled Nanotubes from Carbon Monoxide via the HiPco Process: A Parametric Study”, J. Vac. Sci. Technol., A 2001, 19 (4), 1800–1805. [231] B. Kitiyanan, W. Alvarez, J. Harwell, D. Resasco, “Controlled Production of Single-Wall Carbon Nanotubes by Catalytic Decomposition of CO on Bimetallic Co-Mo Catalysts”, Chem. Phys. Lett. 2000, 317 (3-5), 497–503. [232] P. Nikolaev, M. J. Bronikowski, R. Bradley, F. Rohmund, D. T. Colbert, K. Smith, R. E. Smalley, “Gas-Phase Catalytic Growth of Single-Walled Carbon Nanotubes from Carbon Monoxide”, Chem. Phys. Lett 1999, 313 (12), 91–97. [233] M. Kumar, Y. Ando, “Chemical Vapor Deposition of Carbon Nanotubes: A Review on Growth Mechanism and Mass Production”, J. Nanosci. Nanotechnol. 2010, 10 (6), 3739–3758.

175

Bibliography [234] K. Lu, R. Lago, Y. Chen, M. Green, P. Harris, S. Tsang, “Mechanical Damage of Carbon Nanotubes by Ultrasound”, Carbon 1996, 34 (6), 814–816. [235] M. J. O’Connell, S. M. Bachilo, C. B. Huffman, V. C. Moore, M. S. Strano, E. H. Haroz, K. L. Rialon, P. J. Boul, W. H. Noon, C. Kittrell, J. Ma, R. H. Hauge, R. B. Weisman, R. E. Smalley, “Band Gap Fluorescence from Individual Single-Walled Carbon Nanotubes”, Science 2002, 297 (5581), 593–596. [236] K. Yurekli, C. A. Mitchell, R. Krishnamoorti, “Small-Angle Neutron Scattering from Surfactant-Assisted Aqueous Dispersions of Carbon Nanotubes”, J. Am. Chem. Soc. 2004, 126 (32), 9902–9903. [237] C. Richard, F. Balavoine, P. Schultz, T. W. Ebbesen, C. Mioskowski, “Supramolecular Self-Assembly of Lipid Derivatives on Carbon Nanotubes”, Science 2003, 300 (5620), 775–778. [238] B. Vigolo, A. P´enicaud, C. Coulon, C. Sauder, R. Pailler, C. Journet, P. Bernier, P. Poulin, “Macroscopic Fibers and Ribbons of Oriented Carbon Nanotubes”, Science 2000, 290 (5495), 1331–1334. [239] T. Hertel, A. Hagen, V. Talalaev, K. Arnold, F. Hennrich, M. Kappes, S. Rosenthal, J. McBride, H. Ulbricht, E. Flahaut, “Spectroscopy of Singleand Double-Wall Carbon Nanotubes in Different Environments”, Nano Lett. 2005, 5 (3), 511–514. [240] M. Zheng, A. Jagota, E. D. Semke, B. A. Diner, R. S. Mclean, S. R. Lustig, R. E. Richardson, N. G. Tassi, “DNA-Assisted Dispersion and Separation of Carbon Nanotubes”, Nat. Nanotechnol. 2003, 2 (5), 338–342. [241] M. J. O’Connell, P. Boul, L. M. Ericson, C. Huffman, Y. Wang, E. Haroz, C. Kuper, J. Tour, K. D. Ausman, R. E. Smalley, “Reversible WaterSolubilization of Single-Walled Carbon Nanotubes by Polymer Wrapping”, Chem. Phys. Lett. 2001, 342 (3-4), 265–271. [242] S. A. Curran, P. M. Ajayan, W. J. Blau, D. L. Carroll, J. N. Coleman, A. B. Dalton, A. P. Davey, A. Drury, B. McCarthy, S. Maier, A. Strevens, “A Composite from Poly(m-Phenylenevinylene-co-2,5-Dioctoxy-p-Phenylenevinylene) and Carbon Nanotubes: A Novel Material for Molecular Optoelectronics”, Adv. Mater. 1998, 10 (14), 1091–1093. [243] A. B. Dalton, C. Stephan, J. N. Coleman, B. McCarthy, P. M. Ajayan, S. Lefrant, P. Bernier, W. J. Blau, H. J. Byrne, “Selective Interaction of a Semiconjugated Organic Polymer with Single-Wall Nanotubes”, J. Phys. Chem. B 2000, 104 (43), 10012–10016. [244] A. Nish, J.-Y. Hwang, J. Doig, R. J. Nicholas, “Highly Selective Dispersion of Single-Walled Carbon Nanotubes Using Aromatic Polymers”, Nat. Nanotechnol. 2007, 2 (10), 640–646.

176

Bibliography [245] K. D. Ausman, R. Piner, O. Lourie, R. S. Ruoff, M. Korobov, “Organic Solvent Dispersions of Single-Walled Carbon Nanotubes: Toward Solutions of Pristine Nanotubes”, J. Phys. Chem. B 2000, 104 (38), 8911–8915. [246] F. Bonaccorso, P.-H. Tan, A. C. Ferrari, “Multiwall Nanotubes, Multilayers, and Hybrid Nanostructures: New Frontiers for Technology and Raman Spectroscopy”, ACS Nano 2013, 7 (3), 1838–1844. [247] M. S. Arnold, A. A. Green, J. F. Hulvat, S. I. Stupp, M. C. Hersam, “Sorting Carbon Nanotubes by Electronic Structure Using Density Differentiation”, Nat. Nanotechnol. 2006, 1 (1), 60–65. [248] M. C. Hersam, “Progress towards Monodisperse Single-Walled Carbon Nanotubes”, Nat. Nanotechnol. 2008, 3 (7), 387–394. [249] C. N. R. Rao, R. Voggu, A. Govindaraj, “Selective Generation of SingleWalled Carbon Nanotubes with Metallic, Semiconducting and other Unique Electronic Properties”, Nanoscale 2009, 1 (1), 96–105. [250] H. Liu, T. Tanaka, Y. Urabe, H. Kataura, “High-Efficiency Single-Chirality Separation of Carbon Nanotubes Using Temperature-Controlled Gel Chromatography”, Nano Lett. 2013, 13 (5), 1996–2003. [251] R. Krupke, F. Hennrich, H. v. L¨ohneysen, M. M. Kappes, “Separation of Metallic from Semiconducting Single-Walled Carbon Nanotubes”, Science 2003, 301 (5631), 344–347. [252] S. Ghosh, S. M. Bachilo, R. B. Weisman, “Advanced Sorting of Single-Walled Carbon Nanotubes by Nonlinear Density-Gradient Ultracentrifugation”, Nat. Nanotechnol. 2010, 5 (6), 443–450. [253] Y. Maeda, S.-i. Kimura, M. Kanda, Y. Hirashima, T. Hasegawa, T. Wakahara, Y. Lian, T. Nakahodo, T. Tsuchiya, T. Akasaka, J. Lu, X. Zhang, , Y. Yu, S. Nagase, S. Kazaoui, N. Minami, T. Shimizu, H. Tokumoto, R. Saito, “Large-Scale Separation of Metallic and Semiconducting Single-Walled Carbon Nanotubes”, J. Am. Chem. Soc. 2005, 127 (29), 10287–10290. [254] C. Fantini, A. Jorio, M. Souza, M. Strano, M. Dresselhaus, M. Pimenta, “Optical Transition Energies for Carbon Nanotubes from Resonant Raman Spectroscopy: Environment and Temperature Effects”, Phys. Rev. Lett. 2004, 93 (14), 147406. [255] S. M. Bachilo, M. S. Strano, C. Kittrell, R. H. Hauge, R. E. Smalley, R. B. Weisman, “Structure-Assigned Optical Spectra of Single-Walled Carbon Nanotubes”, Science 2002, 298 (5602), 2361–2366. [256] A. Jorio, M. A. Pimenta, A. G. S. Filho, R. Saito, G. Dresselhaus, M. S. Dresselhaus, “Characterizing Carbon Nanotube Samples with Resonance Raman Scattering”, New J. Phys. 2003, 5 (1), 139.

177

Bibliography [257] H. Kataura, Y. Kumazawa, Y. Maniwa, I. Umezu, S. Suzuki, Y. Ohtsuka, Y. Achiba, “Optical Properties of Single-Wall Carbon Nanotubes”, Synth. Met. 1999, 103 (1-3), 2555–2558. [258] S. Reich, C. Thomsen, P. Orde´on, “Electronic Band Structure of Isolated and Bundled Carbon Nanotubes”, Phys. Rev. B 2002, 65 (15), 155411. [259] G. Wedler, Lehrbuch der Physikalische Chemie, Bd. 5. Auflage, WILEY-VCH Verlag GmbH & Co. KGaA, 2004. [260] A. Jorio, R. Saito, J. Hafner, C. Lieber, M. Hunter, T. McClure, G. Dresselhaus, M. Dresselhaus, “Structural (n, m) Determination of Isolated SingleWall Carbon Nanotubes by Resonant Raman Scattering”, Phys. Rev. Lett. 2001, 86 (6), 1118–1121. ¨ u, [261] A. Jorio, M. Pimenta, A. Souza Filho, G. Samsonidze, A. Swan, M. Unl¨ B. Goldberg, R. Saito, G. Dresselhaus, M. Dresselhaus, “Resonance Raman Spectra of Carbon Nanotubes by Cross-Polarized Light”, Phys. Rev. Lett. 2003, 90 (10), 107403. [262] C. T. White, T. N. Todorov, “Carbon Nanotubes as Long Ballistic Conductors”, Nature 1998, 393 (6682), 240–242. [263] T. D¨ urkop, S. A. Getty, E. Cobas, M. S. Fuhrer, “Extraordinary Mobility in Semiconducting Carbon Nanotubes”, Nano Lett. 2004, 4 (1), 35–39. [264] J.-C. Charlier, X. Blase, S. Roche, “Electronic and Transport Properties of Nanotubes”, Rev. Mod. Phys. 2007, 79 (2), 677–732. [265] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva, A. A. Firsov, “Electric Field Effect in Atomically Thin Carbon Films”, Science 2004, 306 (5696), 666–669. [266] P. Wallace, “The Band Theory of Graphite”, Phys. Rev. 1947, 71 (9), 622– 634. [267] J. Slonczewski, P. Weiss, “Band Structure of Graphite”, Phys. Rev. 1958, 109 (2), 272–279. [268] K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, V. V. Khotkevich, S. V. Morozov, A. K. Geim, “Two-Dimensional Atomic Crystals”, Proc. Natl. Acad. Sci. U. S. A. 2005, 102 (30), 10451–10453. [269] S. Banerjee, M. Sardar, N. Gayathri, A. K. Tyagi, B. Raj, “Enhanced Conductivity in Graphene Layers and at their Edges”, Appl. Phys. Lett. 2006, 88 (6), 062111. [270] A. K. Geim, K. S. Novoselov, “The Rise of Graphene”, Nat. Mater. 2007, 6 (3), 183–191.

178

Bibliography [271] B. Partoens, F. Peeters, “From Graphene to Graphite: Electronic Structure around the K Point”, Phys. Rev. B 2006, 74 (7), 075404. [272] M. D. Stoller, S. Park, Y. Zhu, J. An, R. S. Ruoff, “Graphene-Based Ultracapacitors”, Nano Lett. 2008, 8 (10), 3498–3502. [273] R. R. Nair, P. Blake, A. N. Grigorenko, K. S. Novoselov, T. J. Booth, T. Stauber, N. M. R. Peres, A. K. Geim, “Fine Structure Constant Defines Visual Transparency of Graphene”, Science 2008, 320 (5881), 1308. [274] A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, C. N. Lau, “Superior Thermal Conductivity of Single-Layer Graphene”, Nano Lett. 2008, 8 (3), 902–907. [275] C. Lee, X. Wei, J. W. Kysar, J. Hone, “Measurement of the Elastic Properties and Intrinsic Strength of Monolayer Graphene”, Science 2008, 321 (5887), 385–388. [276] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, M. I. Katsnelson, I. V. Grigorieva, S. V. Dubonos, A. A. Firsov, “Two-Dimensional Gas of Massless Dirac Fermions in Graphene”, Nature 2005, 438 (7065), 197–200. [277] Y. Zhang, Y.-W. Tan, H. L. Stormer, P. Kim, “Experimental Observation of the Quantum Hall Effect and Berry’s Phase in Graphene”, Nature 2005, 438 (7065), 201–204. [278] X. Du, I. Skachko, A. Barker, E. Y. Andrei, “Approaching Ballistic Transport in Suspended Graphene”, Nat. Nanotechnol. 2008, 3 (8), 491–495. [279] A. S. Mayorov, R. V. Gorbachev, S. V. Morozov, L. Britnell, R. Jalil, L. A. Ponomarenko, P. Blake, K. S. Novoselov, K. Watanabe, T. Taniguchi, A. K. Geim, “Micrometer-Scale Ballistic Transport in Encapsulated Graphene at Room Temperature”, Nano Lett. 2011, 11 (6), 2396–2399. [280] K. Bolotin, K. Sikes, Z. Jiang, M. Klima, G. Fudenberg, J. Hone, P. Kim, H. Stormer, “Ultrahigh Electron Mobility in Suspended Graphene”, Solid State Commun. 2008, 146 (9-10), 351–355. [281] D. C. Elias, R. V. Gorbachev, A. S. Mayorov, S. V. Morozov, A. A. Zhukov, P. Blake, L. A. Ponomarenko, I. V. Grigorieva, K. S. Novoselov, F. Guinea, A. K. Geim, “Dirac Cones Reshaped by Interaction Effects in Suspended Graphene”, Nat. Phys. 2011, 7 (9), 701–704. [282] C. Berger, Z. Song, X. Li, X. Wu, N. Brown, C. Naud, D. Mayou, T. Li, J. Hass, A. N. Marchenkov, E. H. Conrad, P. N. First, W. A. de Heer, “Electronic Confinement and Coherence in Patterned Epitaxial Graphene”, Science 2006, 312 (5777), 1191–1196.

179

Bibliography [283] K. V. Emtsev, A. Bostwick, K. Horn, J. Jobst, G. L. Kellogg, L. Ley, J. L. McChesney, T. Ohta, S. A. Reshanov, J. Rohrl, E. Rotenberg, A. K. Schmid, D. Waldmann, H. B. Weber, T. Seyller, “Towards Wafer-Size Graphene Layers By Atmospheric Pressure Graphitization of Silicon Carbide”, Nat. Mater. 2009, 8 (3), 203–207. [284] X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee, L. Colombo, R. S. Ruoff, “Large-Area Synthesis of High-Quality and Uniform Graphene Films on Copper Foils”, Science 2009, 324 (5932), 1312–1314. [285] A. Reina, X. Jia, J. Ho, D. Nezich, H. Son, V. Bulovi´c, M. S. Dresselhaus, J. Kong, “Large Area, Few-Layer Graphene Films on Arbitrary Substrates by Chemical Vapor Deposition”, Nano Lett. 2009, 9 (1), 30–35. [286] K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, J.-H. Ahn, P. Kim, J.-Y. Choi, B. H. Hong, “Large-Scale Pattern Growth of Graphene Films for Stretchable Transparent Electrodes”, Nature 2009, 457 (7230), 706–710. [287] D. W. Boukhvalov, M. I. Katsnelson, “Modeling of Graphite Oxide”, J. Am. Chem. Soc. 2008, 130 (32), 10697–10701. [288] G. Eda, G. Fanchini, M. Chhowalla, “Large-Area Ultrathin Films of Reduced Graphene Oxide as a Transparent and Flexible Electronic Material”, Nat. Nanotechnol. 2008, 3 (5), 270–274. [289] H. Chang, G. Wang, A. Yang, X. Tao, X. Liu, Y. Shen, Z. Zheng, “A Transparent, Flexible, Low-Temperature, and Solution-Processible Graphene Composite Electrode”, Adv. Funct. Mater. 2010, 20 (17), 2893–2902. [290] Y. Hernandez, V. Nicolosi, M. Lotya, F. M. Blighe, Z. Sun, S. De, M. T., B. Holland, M. Byrne, Y. K. Gun’Ko, J. J. Boland, P. Niraj, G. Duesberg, S. Krishnamurthy, R. Goodhue, J. Hutchison, V. Scardaci, A. C. Ferrari, J. N. Coleman, “High-Yield Production of Graphene by Liquid-Phase Exfoliation of Graphite”, Nat. Nanotechnol. 2008, 3 (9), 563–568. [291] P. Blake, P. D. Brimicombe, R. R. Nair, T. J. Booth, D. Jiang, F. Schedin, L. A. Ponomarenko, S. V. Morozov, H. F. Gleeson, E. W. Hill, A. K. Geim, K. S. Novoselov, “Graphene-Based Liquid Crystal Device”, Nano Lett. 2008, 8 (6), 1704–1708. [292] C. E. Hamilton, J. R. Lomeda, Z. Sun, J. M. Tour, A. R. Barron, “High-Yield Organic Dispersions of Unfunctionalized Graphene”, Nano Lett. 2009, 9 (10), 3460–3462. [293] K. S. Novoselov, V. I. Fal’ko, L. Colombo, P. R. Gellert, M. G. Schwab, K. Kim, “A Roadmap for Graphene”, Nature 2012, 490 (7419), 192–200.

180

Bibliography [294] M. Lotya, Y. Hernandez, P. J. King, R. J. Smith, V. Nicolosi, L. S. Karlsson, F. M. Blighe, S. De, Z. Wang, I. T. McGovern, G. S. Duesberg, J. N. Coleman, “Liquid Phase Production of Graphene by Exfoliation of Graphite in Surfactant/Water Solutions”, J. Am. Chem. Soc. 2009, 131 (10), 3611– 3620. [295] J. Cai, P. Ruffieux, R. Jaafar, M. Bieri, T. Braun, S. Blankenburg, M. Muoth, A. P. Seitsonen, M. Saleh, X. Feng, K. Mullen, R. Fasel, “Atomically Precise Bottom-Up Fabrication of Graphene Nanoribbons”, Nature 2010, 466 (7305), 470–473. [296] C.-Y. Su, A.-Y. Lu, Y. Xu, F.-R. Chen, A. N. Khlobystov, L.-J. Li, “HighQuality Thin Graphene Films from Fast Electrochemical Exfoliation”, ACS Nano 2011, 5 (3), 2332–2339. [297] A. Ferrari, J. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. Novoselov, S. Roth, A. Geim, “Raman Spectrum of Graphene and Graphene Layers”, Phys. Rev. Lett. 2006, 97 (18), 187401. [298] L. Malard, M. Pimenta, G. Dresselhaus, M. Dresselhaus, “Raman Spectroscopy in Graphene”, Phys. Rep. 2009, 473 (5-6), 51–87. [299] D. Graf, F. Molitor, K. Ensslin, C. Stampfer, A. Jungen, C. Hierold, L. Wirtz, “Spatially Resolved Raman Spectroscopy of Single- and Few-Layer Graphene”, Nano Lett. 2007, 7 (2), 238–242. [300] A. D. Yoffe, “Electronic Properties of some Chain and Layer Compounds”, Chem. Soc. Rev. 1976, 5 (0), 51–78. [301] P. Joensen, R. Frindt, S. Morrison, “Single-Layer MoS2 ”, Mater. Res. Bull. 1986, 21 (4), 457–461. [302] D. Kong, H. Wang, J. J. Cha, M. Pasta, K. J. Koski, J. Yao, Y. Cui, “Synthesis of MoS2 and MoSe2 Films with Vertically Aligned Layers”, Nano Lett. 2013, 13 (3), 1341–1347. [303] N. Wakabayashi, H. Smith, R. Nicklow, “Lattice Dynamics of Hexagonal MoS2 Studied by Neutron Scattering”, Phys. Rev. B 1975, 12 (2), 659–663. [304] C. Lee, H. Yan, L. E. Brus, T. F. Heinz, J. Hone, S. Ryu, “Anomalous Lattice Vibrations of Single- and Few-Layer MoS2 ”, ACS Nano 2010, 4 (5), 2695–2700. [305] B. Radisavljevic, A. Kis, “Mobility Engineering and a Metal-Insulator Transition in Monolayer MoS2 ”, Nat. Mater. 2013, 12 (9), 815–820. [306] H. R. Guti´errez, N. Perea-L´opez, A. L. El´ıas, A. Berkdemir, B. Wang, R. Lv, F. L´opez-Ur´ıas, V. H. Crespi, H. Terrones, M. Terrones, “Extraordinary Room-Temperature Photoluminescence in Triangular WS2 Monolayers”, Nano Lett. 2013, 13 (8), 3447–3454.

181

Bibliography [307] W. Schutte, J. D. Boer, F. Jellinek, “Crystal Structures of Tungsten Disulfide and Diselenide”, J. Solid State Chem. 1987, 70 (2), 207–209. [308] K. Mak, C. Lee, J. Hone, J. Shan, T. Heinz, “Atomically Thin MoS2 : A New Direct-Gap Semiconductor”, Phys. Rev. Lett. 2010, 105 (13), 136805. [309] M. Chhowalla, H. S. Shin, G. Eda, L.-J. Li, K. P. Loh, H. Zhang, “The Chemistry of Two-Dimensional Layered Transition Metal Dichalcogenide Nanosheets”, Nat. Chem. 2013, 5 (4), 263–275. [310] K. K. Kam, B. A. Parkinson, “Detailed Photocurrent Spectroscopy of the Semiconducting Group VIB Transition Metal Dichalcogenides”, J. Phys. Chem. 1982, 86 (4), 463–467. [311] A. Kuc, N. Zibouche, T. Heine, “Influence of Quantum Confinement on the Electronic Structure of the Transition Metal Sulfide TS2 ”, Phys. Rev. B 2011, 83 (24), 245213. [312] A. Splendiani, L. Sun, Y. Zhang, T. Li, J. Kim, C.-Y. Chim, G. Galli, F. Wang, “Emerging Photoluminescence in Monolayer MoS2 ”, Nano Lett. 2010, 10 (4), 1271–1275. [313] A. L. El´ıas, N. Perea-L´opez, A. Castro-Beltr´an, A. Berkdemir, R. Lv, S. Feng, A. D. Long, T. Hayashi, Y. A. Kim, M. Endo, H. R. Guti´errez, N. R. Pradhan, L. Balicas, T. E. Mallouk, F. L´opez-Ur´ıas, H. Terrones, M. Terrones, “Controlled Synthesis and Transfer of Large-Area WS2 Sheets: From Single Layer to Few Layers”, ACS Nano 2013, 7 (6), 5235–5242. [314] G. Frey, S. Elani, M. Homyonfer, Y. Feldman, R. Tenne, “Optical-Absorption Spectra of Inorganic Fullerenelike MS2 (M = Mo, W)”, Phys. Rev. B 1998, 57 (11), 6666–6671. [315] G. L. Frey, R. Tenne, M. J. Matthews, M. S. Dresselhaus, G. Dresselhaus, “Optical Properties of MS2 (M = Mo, W) Inorganic Fullerenelike and Nanotube Material Optical Absorption and Resonance Raman Measurements”, J. Mater. Res. 1998, 13 (9), 2412–2417. [316] S. Leb`egue, O. Eriksson, “Electronic Structure of Two-Dimensional Crystals from Ab Initio Theory”, Phys. Rev. B 2009, 79 (11), 115409. [317] W. Zhao, Z. Ghorannevis, L. Chu, M. Toh, C. Kloc, P.-H. Tan, G. Eda, “Evolution of Electronic Structure in Atomically Thin Sheets of WS2 and WSe2 ”, ACS Nano 2013, 7 (1), 791–797. [318] B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti, A. Kis, “SingleLayer MoS2 Transistors”, Nat. Nanotechnol. 2011, 6 (3), 147–150. [319] H. Liu, P. Ye, “MoS2 Dual-Gate MOSFET With Atomic-Layer-Deposited Al2 O3 as Top-Gate Dielectric”, IEEE Electron Device Lett. 2012, 33 (4), 546–548.

182

Bibliography [320] S. Das, H.-Y. Chen, A. V. Penumatcha, J. Appenzeller, “High Performance Multilayer MoS2 Transistors with Scandium Contacts”, Nano Lett. 2013, 13 (1), 100–105. [321] S. Jo, N. Ubrig, H. Berger, A. B. Kuzmenko, A. F. Morpurgo, “Mono- and Bilayer WS2 Light-Emitting Transistors”, Nano Lett. 2014, 14 (4), 2019– 2025. [322] M. B. Dines, “Lithium Intercalation via n-Butyllithium of the Layered Transition Metal Dichalcogenides”, Mater. Res. Bull. 1975, 10 (4), 287–291. [323] G. Eda, H. Yamaguchi, D. Voiry, T. Fujita, M. Chen, M. Chhowalla, “Photoluminescence from Chemically Exfoliated MoS2 ”, Nano Lett. 2011, 11 (12), 5111–5116. [324] Z. Zeng, Z. Yin, X. Huang, H. Li, Q. He, G. Lu, F. Boey, H. Zhang, “Single-Layer Semiconducting Nanosheets: High-Yield Preparation and Device Fabrication”, Angew. Chem., Int. Ed. 2011, 50 (47), 11093–11097. [325] B. K. Miremadi, S. R. Morrison, “The Intercalation and Exfoliation of Tungsten Disulfide”, J. Appl. Phys. 1988, 63 (10), 4970–4974. [326] R. Gordon, D. Yang, E. Crozier, D. Jiang, R. Frindt, “Structures of exfoliated single layers of WS2 , MoS2 , and MoSe2 in aqueous suspension”, Phys. Rev. B 2002, 65 (12), 125407. [327] D. Voiry, H. Yamaguchi, J. Li, R. Silva, D. C. B. Alves, T. Fujita, M. Chen, T. Asefa, V. B. Shenoy, G. Eda, M. Chhowalla, “Enhanced Catalytic Activity in Strained Chemically Exfoliated WS2 Nanosheets for Hydrogen Evolution”, Nat. Mater. 2013, 12 (9), 850–855. [328] H. S. S. Ramakrishna Matte, A. Gomathi, A. K. Manna, D. J. Late, R. Datta, S. K. Pati, C. N. R. Rao, “MoS2 and WS2 Analogues of Graphene”, Angew. Chem. 2010, 122 (24), 4153–4156. [329] A. Ayari, E. Cobas, O. Ogundadegbe, M. S. Fuhrer, “Realization and Electrical Characterization of Ultrathin Crystals of Layered Transition-Metal Dichalcogenides”, J. Appl. Phys. 2007, 101 (1), 014507. [330] J. N. Coleman, M. Lotya, A. O’Neill, S. D. Bergin, P. J. King, U. Khan, K. Young, A. Gaucher, S. De, R. J. Smith, I. V. Shvets, S. K. Arora, G. Stanton, H.-Y. Kim, K. Lee, G. T. Kim, G. S. Duesberg, T. Hallam, J. J. Boland, J. J. Wang, J. F. Donegan, J. C. Grunlan, G. Moriarty, A. Shmeliov, R. J. Nicholls, J. M. Perkins, E. M. Grieveson, K. Theuwissen, D. W. McComb, P. D. Nellist, V. Nicolosi, “Two-Dimensional Nanosheets Produced by Liquid Exfoliation of Layered Materials”, Science 2011, 331 (6017), 568– 571.

183

Bibliography [331] R. J. Smith, P. J. King, M. Lotya, C. Wirtz, U. Khan, S. De, A. O’Neill, G. S. Duesberg, J. C. Grunlan, G. Moriarty, J. Chen, J. Wang, A. I. Minett, V. Nicolosi, J. N. Coleman, “Large-Scale Exfoliation of Inorganic Layered Compounds in Aqueous Surfactant Solutions”, Adv. Mater. 2011, 23 (34), 3944–3948. [332] P. May, U. Khan, J. M. Hughes, J. N. Coleman, “Role of Solubility Parameters in Understanding the Steric Stabilization of Exfoliated Two-Dimensional Nanosheets by Adsorbed Polymers”, J. Phys. Chem. C 2012, 116 (20), 11393–11400. [333] Q. H. Wang, K. Kalantar-Zadeh, A. Kis, J. N. Coleman, M. S. Strano, “Electronics and Optoelectronics of Two-Dimensional Transition Metal Dichalcogenides”, Nat. Nanotechnol. 2012, 7 (11), 699–712. [334] G. Cunningham, M. Lotya, C. S. Cucinotta, S. Sanvito, S. D. Bergin, R. Menzel, M. S. P. Shaffer, J. N. Coleman, “Solvent Exfoliation of Transition Metal Dichalcogenides: Dispersibility of Exfoliated Nanosheets Varies Only Weakly between Compounds”, ACS Nano 2012, 6 (4), 3468–3480. [335] Y.-H. Lee, X.-Q. Zhang, W. Zhang, M.-T. Chang, C.-T. Lin, K.-D. Chang, Y.-C. Yu, J. T.-W. Wang, C.-S. Chang, L.-J. Li, T.-W. Lin, “Synthesis of Large-Area MoS2 Atomic Layers with Chemical Vapor Deposition”, Adv. Mater. 2012, 24 (17), 2320–2325. [336] Y. Zhan, Z. Liu, S. Najmaei, P. M. Ajayan, J. Lou, “Large-Area Vapor-Phase Growth and Characterization of MoS2 Atomic Layers on a SiO2 Substrate”, Small 2012, 8 (7), 966–971. [337] K.-K. Liu, W. Zhang, Y.-H. Lee, Y.-C. Lin, M.-T. Chang, C.-Y. Su, C.-S. Chang, H. Li, Y. Shi, H. Zhang, C.-S. Lai, L.-J. Li, “Growth of Large-Area and Highly Crystalline MoS2 Thin Layers on Insulating Substrates”, Nano Lett. 2012, 12 (3), 1538–1544. [338] Y. Shi, W. Zhou, A.-Y. Lu, W. Fang, Y.-H. Lee, A. L. Hsu, S. M. Kim, K. K. Kim, H. Y. Yang, L.-J. Li, J.-C. Idrobo, J. Kong, “van der Waals Epitaxy of MoS2 Layers Using Graphene As Growth Templates”, Nano Lett. 2012, 12 (6), 2784–2791. [339] A. M. van der Zande, P. Y. Huang, D. A. Chenet, T. C. Berkelbach, Y. You, G.-H. Lee, T. F. Heinz, D. R. Reichman, D. A. Muller, J. C. Hone, “Grains and Grain Boundaries in Highly Crystalline Monolayer Molybdenum Disulphide”, Nat. Mater. 2013, 12 (6), 554–561. [340] S. Najmaei, Z. Liu, W. Zhou, X. Zou, G. Shi, S. Lei, B. I. Yakobson, J.-C. Idrobo, P. M. Ajayan, J. Lou, “Vapour Phase Growth and Grain Boundary Structure of Molybdenum Disulphide Atomic Layers”, Nat. Mater. 2013, 12 (8), 754–759.

184

Bibliography [341] X. Wang, H. Feng, Y. Wu, L. Jiao, “Controlled Synthesis of Highly Crystalline MoS2 Flakes by Chemical Vapor Deposition”, J. Am. Chem. Soc. 2013, 135 (14), 5304–5307. [342] H. Schmidt, S. Wang, L. Chu, M. Toh, R. Kumar, W. Zhao, A. H. Castro Neto, ¨ J. Martin, S. Adam, B. Ozyilmaz, G. Eda, “Transport Properties of Monolayer MoS2 Grown by Chemical Vapor Deposition”, Nano Lett. 2014, 14 (4), 1909– 1913. [343] M. Shanmugam, T. Bansal, C. A. Durcan, B. Yu, “Schottky-Barrier Solar Cell Based on Layered Semiconductor Tungsten Disulfide Nanofilm”, Appl. Phys. Lett. 2012, 101 (26), 263902. [344] C. Cong, J. Shang, X. Wu, B. Cao, N. Peimyoo, C. Qiu, L. Sun, T. Yu, “Synthesis and Optical Properties of Large-Area Single-Crystalline 2D Semiconductor WS2 Monolayer from Chemical Vapor Deposition”, Adv. Opt. Mater. 2014, 2 (2), 131–136. [345] N. Perea-L´opez, A. L. El´ıas, A. Berkdemir, A. Castro-Beltran, H. R. Guti´errez, S. Feng, R. Lv, T. Hayashi, F. L´opez-Ur´ıas, S. Ghosh, B. Muchharla, S. Talapatra, H. Terrones, M. Terrones, “Photosensor Device Based on Few-Layered WS2 Films”, Adv. Funct. Mater. 2013, 23 (44), 5511–5517. [346] Y. Zhang, Y. Zhang, Q. Ji, J. Ju, H. Yuan, J. Shi, T. Gao, D. Ma, M. Liu, Y. Chen, X. Song, H. Y. Hwang, Y. Cui, Z. Liu, “Controlled Growth of High-Quality Monolayer WS2 Layers on Sapphire and Imaging Its Grain Boundary”, ACS Nano 2013, 7 (10), 8963–8971. [347] H. Li, Q. Zhang, C. C. R. Yap, B. K. Tay, T. H. T. Edwin, A. Olivier, D. Baillargeat, “From Bulk to Monolayer MoS2 : Evolution of Raman Scattering”, Adv. Funct. Mater. 2012, 22 (7), 1385–1390. [348] D. V. Talapin, J.-S. Lee, M. V. Kovalenko, E. V. Shevchenko, “Prospects of Colloidal Nanocrystals for Electronic and Optoelectronic Applications”, Chem. Rev. 2010, 110 (1), 389–458. [349] H. S. Lee, S.-W. Min, Y.-G. Chang, M. K. Park, T. Nam, H. Kim, J. H. Kim, S. Ryu, S. Im, “MoS2 Nanosheet Phototransistors with Thickness-Modulated Optical Energy Gap”, Nano Lett. 2012, 12 (7), 3695–3700. [350] D. Eder, “Carbon Nanotube-Inorganic Hybrids”, Chem. Rev. 2010, 110 (3), 1348–1385. [351] X. Huang, X. Qi, F. Boey, H. Zhang, “Graphene-based Composites”, Chem. Soc. Rev. 2012, 41 (2), 666–686. [352] G. Konstantatos, M. Badioli, L. Gaudreau, J. Osmond, M. Bernechea, F. P. G. de Arquer, F. Gatti, F. H. L. Koppens, “Hybrid Graphene-Quantum Dot Phototransistors with Ultrahigh Gain”, Nat. Nanotechnol. 2012, 7 (6), 363–368.

185

Bibliography [353] C. X. Guo, H. B. Yang, Z. M. Sheng, Z. S. Lu, Q. L. Song, C. M. Li, “Layered Graphene/Quantum Dots for Photovoltaic Devices”, Angew. Chem., Int. Ed. 2010, 49 (17), 3014–3017. [354] T. S. Sreeprasad, V. Berry, “How Do the Electrical Properties of Graphene Change with its Functionalization?”, Small 2013, 9 (3), 341–350. [355] K. K. Manga, J. Wang, M. Lin, J. Zhang, M. Nesladek, V. Nalla, W. Ji, K. P. Loh, “High-Performance Broadband Photodetector Using SolutionProcessible PbSe-TiO2 -Graphene Hybrids”, Adv. Mater. 2012, 24 (13), 1697– 1702. [356] Z. Shi, C. Liu, W. Lv, H. Shen, D. Wang, L. Chen, L. S. Li, J. Jin, “FreeStanding Single-Walled Carbon Nanotube-CdSe Quantum Dots Hybrid Ultrathin Films for Flexible Optoelectronic Conversion Devices”, Nanoscale 2012, 4 (15), 4515–4521. [357] S. Banerjee, S. S. Wong, “Synthesis and Characterization of Carbon Nanotube-Nanocrystal Heterostructures”, Nano Lett. 2002, 2 (3), 195–200. [358] J. M. Haremza, M. A. Hahn, T. D. Krauss, S. Chen, J. Calcines, “Attachment of Single CdSe Nanocrystals to Individual Single-Walled Carbon Nanotubes”, Nano Lett. 2002, 2 (11), 1253–1258. [359] S. Ravindran, S. Chaudhary, B. Colburn, M. Ozkan, C. S. Ozkan, “Covalent Coupling of Quantum Dots to Multiwalled Carbon Nanotubes for Electronic Device Applications”, Nano Lett. 2003, 3 (4), 447–453. [360] L. Sheeney-Haj-Ichia, B. Basnar, I. Willner, “Efficient Generation of Photocurrents by Using CdS/Carbon Nanotube Assemblies on Electrodes”, Angew. Chem., Int. Ed. 2005, 44 (1), 78–83. [361] B. Pan, D. Cui, R. He, F. Gao, Y. Zhang, “Covalent Attachment of Quantum Dot on Carbon Nanotubes”, Chem. Phys. Lett 2006, 417 (4-6), 419–424. [362] B. Zebli, H. Vieyra, I. Carmeli, A. Hartschuh, J. Kotthaus, A. Holleitner, “Optoelectronic Sensitization of Carbon Nanotubes by CdTe Nanocrystals”, Phys. Rev. B 2009, 79 (20), 205402. [363] D. Wang, J. K. Baral, H. Zhao, B. A. Gonfa, V.-V. Truong, M. A. El Khakani, R. Izquierdo, D. Ma, “Controlled Fabrication of PbS Quantum-Dot/CarbonNanotube Nanoarchitecture and its Significant Contribution to Near-Infrared Photon-to-Current Conversion”, Adv. Funct. Mater. 2011, 21 (21), 4010– 4018. [364] N. Cho, K. Roy Choudhury, R. B. Thapa, Y. Sahoo, T. Ohulchanskyy, A. N. Cartwright, K.-S. Lee, P. N. Prasad, “Efficient Photodetection at IR Wavelengths by Incorporation of PbSe-Carbon-Nanotube Conjugates in a Polymeric Nanocomposite”, Adv. Mater. 2007, 19 (2), 232–236.

186

Bibliography [365] L. Hu, Y.-L. Zhao, K. Ryu, C. Zhou, J. F. Stoddart, G. Gr¨ uner, “LightInduced Charge Transfer in Pyrene/CdSe-SWNT Hybrids”, Adv. Mater. 2008, 20 (5), 939–946. [366] J. E. Weaver, M. R. Dasari, A. Datar, S. Talapatra, P. Kohli, “Investigating Photoinduced Charge Transfer in Carbon Nanotube-Perylene-Quantum Dot Hybrid Nanocomposites”, ACS Nano 2010, 4 (11), 6883–6893. [367] D. M. Guldi, G. M. A. Rahman, V. Sgobba, N. A. Kotov, D. Bonifazi, M. Prato, “CNT-CdTe Versatile Donor-Acceptor Nanohybrids”, J. Am. Chem. Soc. 2006, 128 (7), 2315–2323. no Maceira, M. Giersig, [368] M. Grzelczak, M. A. Correa-Duarte, V. Salgueiri˜ R. Diaz, L. M. Liz-Marz´an, “Photoluminescence Quenching Control in Quantum Dot-Carbon Nanotube Composite Colloids Using a Silica-Shell Spacer”, Adv. Mater. 2006, 18 (4), 415–420. [369] S. Leubner, G. Katsukis, D. M. Guldi, “Decorating Polyelectrolyte Wrapped SWNTs with CdTe Quantum Dots for Solar Energy Conversion”, Faraday Discuss. 2012, 155 (297-308), 253–265. [370] M. Olek, T. B¨ usgen, M. Hilgendorff, M. Giersig, “Quantum Dot Modified Multiwall Carbon Nanotubes”, J. Phys. Chem. B 2006, 110 (26), 12901– 12904. [371] X. Li, Y. Jia, A. Cao, “Tailored Single-Walled Carbon Nanotube-CdS Nanoparticle Hybrids for Tunable Optoelectronic Devices”, ACS Nano 2010, 4 (1), 506–512. [372] X. Li, Y. Jia, J. Wei, H. Zhu, K. Wang, D. Wu, A. Cao, “Solar Cells and Light Sensors Based on Nanoparticle-Grafted Carbon Nanotube Films”, ACS Nano 2010, 4 (4), 2142–2148. [373] X. Peng, J. A. Misewich, S. S. Wong, M. Y. Sfeir, “Efficient Charge Separation in Multidimensional Nanohybrids”, Nano Lett. 2011, 11 (11), 4562–4568. [374] I. Ka, V. Le Borgne, D. Ma, M. A. El Khakani, “Pulsed Laser Ablation based Direct Synthesis of Single-Wall Carbon Nanotube/PbS Quantum Dot Nanohybrids Exhibiting Strong, Spectrally Wide and Fast Photoresponse”, Adv. Mater. 2012, 24 (47), 6289–6294. [375] K. Yu, G. Lu, K. Chen, S. Mao, H. Kim, J. Chen, “Controllable Photoelectron Transfer in CdSe Nanocrystal-Carbon Nanotube Hybrid Structures”, Nanoscale 2012, 4 (3), 742–746. [376] S. Banerjee, S. S. Wong, “In Situ Quantum Dot Growth on Multiwalled Carbon Nanotubes”, J. Am. Chem. Soc. 2003, 125 (34), 10342–10350.

187

Bibliography [377] S. Banerjee, S. S. Wong, “In-Situ Growth of “Fused”, Ozonized Single-Walled Carbon Nanotubes-CdTe Quantum Dot Junctions”, Adv. Mater. 2004, 16 (1), 34–37. [378] S. Banerjee, S. S. Wong, “Formation of CdSe Nanocrystals onto Oxidized, Ozonized Single-Walled Carbon Nanotube Surfaces”, Chem. Commun. 2004, (16), 1866–1867. [379] I. Robel, B. A. Bunker, P. V. Kamat, “Single-Walled Carbon Nanotube-CdS Nanocomposites as Light-Harvesting Assemblies: Photoinduced ChargeTransfer Interactions”, Adv. Mater. 2005, 17 (20), 2458–2463. [380] B. H. Ju´arez, C. Klinke, A. Kornowski, H. Weller, “Quantum Dot Attachment and Morphology Control by Carbon Nanotubes”, Nano Lett. 2007, 7 (12), 3564–3568. [381] A. B. Hungr´ıa, B. H. Ju´arez, C. Klinke, H. Weller, P. A. Midgley, “3-D Characterization of CdSe Nanoparticles Attached to Carbon Nanotubes”, Nano Res. 2008, 1 (1), 89–97. [382] D. Yu, Y. Chen, B. Li, X. Chen, M. Zhang, “Fabrication and Characterization of PbS/Multiwalled Carbon Nanotube Heterostructures”, Appl. Phys. Lett. 2007, 90 (16), 161103. [383] X. Geng, L. Niu, Z. Xing, R. Song, G. Liu, M. Sun, G. Cheng, H. Zhong, Z. Liu, Z. Zhang, L. Sun, H. Xu, L. Lu, L. Liu, “Aqueous-Processable Noncovalent Chemically Converted Graphene-Quantum Dot Composites for Flexible and Transparent Optoelectronic Films”, Adv. Mater. 2010, 22 (5), 638–642. [384] S. Guo, D. Bao, S. Upadhyayula, W. Wang, A. B. Guvenc, J. R. Kyle, H. Hosseinibay, K. N. Bozhilov, V. I. Vullev, C. S. Ozkan, M. Ozkan, “Photoinduced Electron Transfer Between Pyridine Coated Cadmium Selenide Quantum Dots and Single Sheet Graphene”, Adv. Funct. Mater. 2013, 23 (41), 5199–5211. [385] G. Katsukis, J. Malig, C. Schulz-Drost, S. Leubner, N. Jux, D. M. Guldi, “Toward Combining Graphene and QDs: Assembling CdTe QDs to Exfoliated Graphite and Nanographene in Water”, ACS Nano 2012, 6 (3), 1915–1924. [386] Z. Sun, Z. Liu, J. Li, G.-a. Tai, S.-P. Lau, F. Yan, “Infrared Photodetectors Based on CVD-Grown Graphene and PbS Quantum Dots with Ultrahigh Responsivity”, Adv. Mater. 2012, 24 (43), 5878–5883. [387] P. Jing, W. Ji, X. Yuan, M. Ikezawa, L. Zhang, H. Li, J. Zhao, Y. Masumoto, “Photoinduced Charge Separation and Recombination Processes in CdSe Quantum Dot and Graphene Oxide Composites with Methylene Blue as Linker”, J. Phys. Chem. Lett. 2013, 4 (17), 2919–2925.

188

Bibliography [388] A. Cao, Z. Liu, S. Chu, M. Wu, Z. Ye, Z. Cai, Y. Chang, S. Wang, Q. Gong, Y. Liu, “A Facile One-Step Method to Produce Graphene-CdS Quantum Dot Nanocomposites as Promising Optoelectronic Materials”, Adv. Mater. 2010, 22 (1), 103–106. [389] Y. Lin, K. Zhang, W. Chen, Y. Liu, Z. Geng, J. Zeng, N. Pan, L. Yan, X. Wang, J. G. Hou, “Dramatically Enhanced Photoresponse of Reduced Graphene Oxide with Linker-Free Anchored CdSe Nanoparticles”, ACS Nano 2010, 4 (6), 3033–3038. [390] K. Yu, G. Lu, S. Mao, K. Chen, H. Kim, Z. Wen, J. Chen, “Selective Deposition of CdSe Nanoparticles on Reduced Graphene Oxide to Understand Photoinduced Charge Transfer in Hybrid Nanostructures”, ACS Appl. Mater. Interfaces 2011, 3 (7), 2703–2709. [391] Y.-T. Kim, H.-W. Shin, Y.-S. Ko, T. K. Ahn, Y.-U. Kwon, “Synthesis of a CdSe-Graphene Hybrid Composed of CdSe Quantum Dot Arrays Directly Grown on CVD-Graphene and its Ultrafast Carrier Dynamics”, Nanoscale 2013, 5 (4), 1483–1488. [392] D. Zhang, L. Gan, Y. Cao, Q. Wang, L. Qi, X. Guo, “Understanding Charge Transfer at PbS-Decorated Graphene Surfaces toward a Tunable Photosensor”, Adv. Mater. 2012, 24 (20), 2715–2720. [393] Y. Zhang, N. Zhang, Z.-R. Tang, Y.-J. Xu, “Graphene Transforms Wide Band Gap ZnS to a Visible Light Photocatalyst. The New Role of Graphene as a Macromolecular Photosensitizer”, ACS Nano 2012, 6 (11), 9777–9789. [394] D. Kufer, I. Nikitskiy, T. Lasanta, G. Navickaite, F. H. L. Koppens, G. Konstantatos, “Hybrid 2D-0D MoS2 -PbS Quantum Dot Photodetectors”, Adv. Mater. 2015, 27 (1), 176–180. [395] F. Prins, A. J. Goodman, W. A. Tisdale, “Reduced Dielectric Screening and Enhanced Energy Transfer in Single- and Few-Layer MoS2 ”, Nano Lett. 2014, 14 (11), 6087–6091. [396] X. Huang, Z. Zeng, S. Bao, M. Wang, X. Qi, Z. Fan, H. Zhang, “SolutionPhase Epitaxial Growth of Noble Metal Nanostructures on Dispersible SingleLayer Molybdenum Disulfide Nanosheets”, Nat. Commun. 2013, 4, 1444. [397] E. Spicker, M. Garbrecht, W. J¨ager, K. Tillmann, “Advantages of Aberration Correction for HRTEM Investigation of Complex Layer Compounds”, J. Microsc. 2010, 237 (3), 341–346. [398] M. M. Smeller, C. L. Heideman, Q. Lin, M. Beekman, M. D. Anderson, P. Zschack, I. M. Anderson, D. C. Johnson, “Structure of Turbostratically Disordered Misfit Layer Compounds [PbSe0.99 ]1 [WSe2 ]1 , [PbSe1.00 ]1 [MoSe2 ]1 , and [SnSe1.03 ]1 [MoSe2 ]1 ”, Z. anorg. allg. Chem. 2012, 638 (15), 2632–2639.

189

Bibliography [399] G. A. Wiegers, “Misfit Layer Compounds: Structures and Physical Properties”, Prog. Solid State Chem. 1996, 24 (1-2), 1–139. [400] J. Rouxel, A. Meerschaut, G. Wiegers, “Chalcogenide misfit layer compounds”, J. Alloys Compd. 1995, 229 (1), 144–157. [401] J. Brandt, L. Kipp, M. Skibowski, E. Krasovskii, W. Schattke, E. Spiecker, C. Dieker, W. J¨ager, “Charge Transfer in Misfit Layered Compounds”, Surf. Sci. 2003, 532-535, 705–710. [402] E. Kabliman, P. Blaha, K. Schwarz, “Ab Initio Study of Stabilization of the Misfit Layer Compound (PbS)1.14 TaS2 ”, Phys. Rev. B 2010, 82 (12), 125308. [403] S. A. McDonald, G. Konstantatos, S. Zhang, P. W. Cyr, E. J. D. Klem, L. Levina, E. H. Sargent, “Solution-Processed PbS Quantum Dot Infrared Photodetectors and Photovoltaics”, Nat. Mater. 2005, 4 (2), 138–142. [404] I. Gur, N. A. Fromer, M. L. Geier, A. P. Alivisatos, “Air-Stable All-Inorganic Nanocrystal Solar Cells Processed from Solution”, Science 2005, 310 (5747), 462–465. [405] E. J. D. Klem, D. D. MacNeil, P. W. Cyr, L. Levina, E. H. Sargent, “Efficient Solution-Processed Infrared Photovoltaic Cells: Planarized All-Inorganic Bulk Heterojunction Devices via Inter-Quantum-Dot Bridging During Growth from Solution”, Appl. Phys. Lett. 2007, 90 (18), 183113. [406] P. V. Kamat, “Quantum Dot Solar Cells. Semiconductor Nanocrystals as Light Harvesters”, J. Phys. Chem. C 2008, 112 (48), 18737–18753. [407] M. Law, M. C. Beard, S. Choi, J. M. Luther, M. C. Hanna, A. J. Nozik, “Determining the Internal Quantum Efficiency of PbSe Nanocrystal Solar Cells with the Aid of an Optical Model”, Nano Lett. 2008, 8 (11), 3904–3910. [408] J. J. Choi, Y.-F. Lim, M. B. Santiago-Berrios, M. Oh, B.-R. Hyun, L. Sun, A. C. Bartnik, A. Goedhart, G. G. Malliaras, H. D. Abru˜ na, F. W. Wise, T. Hanrath, “PbSe Nanocrystal Excitonic Solar Cells”, Nano Lett. 2009, 9 (11), 3749–3755. [409] N. Zhao, T. P. Osedach, L.-Y. Chang, S. M. Geyer, D. Wanger, M. T. Binda, A. C. Arango, M. G. Bawendi, V. Bulovi´c, “Colloidal PbS Quantum Dot Solar Cells with High Fill Factor”, ACS Nano 2010, 4 (7), 3743–3752. [410] J. Gao, C. L. Perkins, J. M. Luther, M. C. Hanna, H.-Y. Chen, O. E. Semonin, A. J. Nozik, R. J. Ellingson, M. C. Beard, “n-Type Transition Metal Oxide as a Hole Extraction Layer in PbS Quantum Dot Solar Cells”, Nano Lett. 2011, 11 (8), 3263–3266.

190

Bibliography [411] J. Tang, H. Liu, D. Zhitomirsky, S. Hoogland, X. Wang, M. Furukawa, L. Levina, E. H. Sargent, “Quantum Junction Solar Cells”, Nano Lett. 2012, 12 (9), 4889–4894. [412] B.-R. Hyun, J. J. Choi, K. L. Seyler, T. Hanrath, F. W. Wise, “Heterojunction PbS Nanocrystal Solar Cells with Oxide Charge-Transport Layers”, ACS Nano 2013, 7 (12), 10938–10947. [413] C.-H. M. Chuang, P. R. Brown, V. Bulovi´c, M. G. Bawendi, “Improved Performance and Stability in Quantum Dot Solar Cells Through Band Alignment Engineering”, Nat. Mater. 2014, 13 (8), 796–801. [414] B. O. Dabbousi, M. G. Bawendi, O. Onitsuka, M. F. Rubner, “Electroluminescence from CdSe Quantum-Dot/Polymer Composites”, Appl. Phys. Lett. 1995, 66 (11), 1316–1318. [415] M. V. Artemyev, V. Sperling, U. Woggon, “Electroluminescence in Thin Solid Films of Closely Packed CdS Nanocrystals”, J. Appl. Phys. 1997, 81 (10), 6975–6977. [416] M. C. Schlamp, X. Peng, A. P. Alivisatos, “Improved Efficiencies in Light Emitting Diodes Made with CdSe(CdS) Core/Shell Type Nanocrystals and a Semiconducting Polymer”, J. Appl. Phys. 1997, 82 (11), 5837–5842. [417] H. Mattoussi, L. H. Radzilowski, B. O. Dabbousi, E. L. Thomas, M. G. Bawendi, M. F. Rubner, “Electroluminescence from Heterostructures of Poly(Phenylene Vinylene) and Inorganic CdSe Nanocrystals”, J. Appl. Phys. 1998, 83 (12), 7965–7974. [418] N. Tessler, V. Medvedev, M. Kazes, S. Kan, U. Banin, “Efficient Near-Infrared Polymer Nanocrystal Light-Emitting Diodes”, Science 2002, 295 (5559), 1506–1508. [419] L. Bakueva, S. Musikhin, M. A. Hines, T.-W. F. Chang, M. Tzolov, G. D. Scholes, E. H. Sargent, “Size-Tunable Infrared (1000-1600 nm) Electroluminescence from PbS Quantum-Dot Nanocrystals in a Semiconducting Polymer”, Appl. Phys. Lett. 2003, 82 (17), 2895–2897. [420] V. L. Colvin, M. C. Schlamp, A. P. Alivisatos, “Light-Emitting Diodes made from Cadmium Selenide Nanocrystals and a Semiconducting Polymer”, Nature 1994, 370 (6488), 354–357. [421] S. Coe, W.-K. Woo, M. Bawendi, V. Bulovi´c, “Electroluminescence from Single Monolayers of Nanocrystals in Molecular Organic Devices”, Nature 2002, 420 (6917), 800–803. [422] A. L. Rogach, N. Gaponik, J. M. Lupton, C. Bertoni, D. E. Gallardo, S. Dunn, N. Li Pira, M. Paderi, P. Repetto, S. G. Romanov, C. O’Dwyer, C. M. Sotomayor Torres, A. Eychm¨ uller, “Light-Emitting Diodes with Semiconductor Nanocrystals”, Angew. Chem., Int. Ed. 2008, 47 (35), 6538–6549.

191

Bibliography [423] M. Gao, C. Lesser, S. Kirstein, H. Mhwald, A. L. Rogach, H. Weller, “Electroluminescence of Different Colors from Polycation/CdTe Nanocrystal SelfAssembled Films”, J. Appl. Phys. 2000, 87 (5), 2297–2302. [424] J. S. Steckel, P. Snee, S. Coe-Sullivan, J. P. Zimmer, J. E. Halpert, P. Anikeeva, L.-A. Kim, V. Bulovi´c, M. G. Bawendi, “Color-Saturated GreenEmitting QD-LEDs”, Angew. Chem., Int. Ed. 2006, 45 (35), 5796–5799. [425] C. Bertoni, D. Gallardo, S. Dunn, N. Gaponik, A. Eychm¨ uller, “Fabrication and Characterization of Red-Emitting Electroluminescent Devices Based on Thiol-Stabilized Semiconductor Nanocrystals”, Appl. Phys. Lett. 2007, 90 (3), 034107. [426] J.-M. Caruge, J. E. Halpert, V. Bulovi´c, M. G. Bawendi, “NiO as an Inorganic Hole-Transporting Layer in Quantum-Dot Light-Emitting Devices”, Nano Lett. 2006, 6 (12), 2991–2994. [427] A. H. Mueller, M. A. Petruska, M. Achermann, D. J. Werder, E. A. Akhadov, D. D. Koleske, M. A. Hoffbauer, V. I. Klimov, “Multicolor Light-Emitting Diodes Based on Semiconductor Nanocrystals Encapsulated in GaN Charge Injection Layers”, Nano Lett. 2005, 5 (6), 1039–1044. [428] Q. Sun, Y. A. Wang, L. S. Li, D. Wang, T. Zhu, J. Xu, C. Yang, Y. Li, “Bright, Multicoloured Light-Emitting Diodes Based on Quantum Dots”, Nat. Photon. 2007, 1 (12), 717–722. [429] J. Zhao, J. A. Bardecker, A. M. Munro, M. S. Liu, Y. Niu, I.-K. Ding, J. Luo, B. Chen, A. K.-Y. Jen, D. S. Ginger, “Efficient CdSe/CdS Quantum Dot Light-Emitting Diodes Using a Thermally Polymerized Hole Transport Layer”, Nano Lett. 2006, 6 (3), 463–467. [430] S. Chaudhary, M. Ozkan, W. C. W. Chan, “Trilayer Hybrid PolymerQuantum Dot Light-Emitting Diodes”, Appl. Phys. Lett. 2004, 84 (15), 2925–2927. [431] J. Steckel, S. Coe-Sullivan, V. Bulovi´c, M. Bawendi, “1.3 µm to 1.55 µm Tunable Electroluminescence from PbSe Quantum Dots Embedded within an Organic Device”, Adv. Mater. 2003, 15 (21), 1862–1866. [432] P. O. Anikeeva, J. E. Halpert, M. G. Bawendi, V. Bulovi´c, “Quantum Dot Light-Emitting Devices with Electroluminescence Tunable over the Entire Visible Spectrum”, Nano Lett. 2009, 9 (7), 2532–2536. [433] J. S. Bendall, M. Paderi, F. Ghigliotti, N. Li Pira, V. Lambertini, V. Lesnyak, N. Gaponik, G. Visimberga, A. Eychm¨ uller, C. M. S. Torres, M. E. Welland, C. Gieck, L. Marchese, “Layer-by-Layer All-Inorganic Quantum-Dot-Based LEDs: A Simple Procedure with Robust Performance”, Adv. Funct. Mater. 2010, 20 (19), 3298–3302.

192

Bibliography [434] K.-S. Cho, E. K. Lee, W.-J. Joo, E. Jang, T.-H. Kim, S. J. Lee, S.-J. Kwon, J. Y. Han, B.-K. Kim, B. L. Choi, J. M. Kim, “High-Performance Crosslinked Colloidal Quantum-Dot Light-Emitting Diodes”, Nat. Photon. 2009, 3 (6), 341–345. [435] L. Qian, Y. Zheng, J. Xue, P. H. Holloway, “Stable and Efficient Quantum-Dot Light-Emitting Diodes Based on Solution-Processed Multilayer Structures”, Nat. Photon. 2011, 5 (9), 543–548. [436] Y. L. Kong, I. A. Tamargo, H. Kim, B. N. Johnson, M. K. Gupta, T.-W. Koh, H.-A. Chin, D. A. Steingart, B. P. Rand, M. C. McAlpine, “3D Printed Quantum Dot Light-Emitting Diodes”, Nano Lett. 2014, 14 (12), 7017–7023. [437] L. Sun, J. J. Choi, D. Stachnik, A. C. Bartnik, B.-R. Hyun, G. G. Malliaras, T. Hanrath, F. W. Wise, “Bright Infrared Quantum-Dot Light-Emitting Diodes Through Inter-Dot Spacing Control”, Nat. Nanotechnol. 2012, 7 (6), 369–373. [438] G. J. Supran, K. W. Song, G. W. Hwang, R. E. Correa, J. Scherer, E. A. Dauler, Y. Shirasaki, M. G. Bawendi, V. Bulovi´c, “High-Performance Shortwave-Infrared Light-Emitting Devices Using Core-Shell (PbS-CdS) Colloidal Quantum Dots”, Adv. Mater. 2015, 27 (8), 1437–1442. [439] G. Konstantatos, E. H. Sargent, “Solution-Processed Quantum Dot Photodetectors”, Proc. IEEE 2009, 97 (10), 1666–1683. [440] S. M. Sze, K. K. Ng, Physics of Semiconductor Devices, 3. Aufl., John Wiley & Sons, Inc., New York, 2002. [441] G. Konstantatos, J. Clifford, L. Levina, E. H. Sargent, “Sensitive SolutionProcessed Visible-Wavelength Photodetectors”, Nat. Photon. 2007, 1 (9), 531–534. [442] G. Konstantatos, L. Levina, A. Fischer, E. H. Sargent, “Engineering the Temporal Response of Photoconductive Photodetectors via Selective Introduction of Surface Trap States”, Nano Lett. 2008, 8 (5), 1446–1450. [443] N. C. Greenham, X. Peng, A. P. Alivisatos, “Charge Separation and Transport in Conjugated-Polymer/Semiconductor-Nanocrystal Composites Studied by Photoluminescence Quenching and Photoconductivity”, Phys. Rev. B 1996, 54 (24), 17628–17637. [444] D. S. Ginger, N. C. Greenham, “Photoinduced Electron Transfer from Conjugated Polymers to CdSe Nanocrystals”, Phys. Rev. B 1999, 59 (16), 10622–10629. [445] C. A. Leatherdale, C. R. Kagan, N. Y. Morgan, S. A. Empedocles, M. A. Kastner, M. G. Bawendi, “Photoconductivity in CdSe Quantum Dot Solids”, Phys. Rev. B 2000, 62 (4), 2669–2680.

193

Bibliography [446] D. S. Ginger, N. C. Greenham, “Charge Injection and Transport in Films of CdSe Nanocrystals”, J. Appl. Phys. 2000, 87 (3), 1361–1368. [447] N. Y. Morgan, C. A. Leatherdale, M. Drndi´c, M. V. Jarosz, M. A. Kastner, M. Bawendi, “Electronic Transport in Films of Colloidal CdSe Nanocrystals”, Phys. Rev. B 2002, 66 (7), 075339. [448] D. Yu, B. L. Wehrenberg, P. Jha, J. Ma, P. Guyot-Sionnest, “Electronic Transport of n-Type CdSe Quantum Dot Films: Effect of Film Treatment”, J. Appl. Phys. 2006, 99 (10), 104315. [449] V. J. Porter, T. Mentzel, S. Charpentier, M. A. Kastner, M. G. Bawendi, “Temperature-, Gate-, and Photoinduced Conductance of Close-Packed CdTe Nanocrystal Films”, Phys. Rev. B 2006, 73 (15), 155303. [450] S. Hinds, L. Levina, E. J. D. Klem, G. Konstantatos, V. Sukhovatkin, E. H. Sargent, “Smooth-Morphology Ultrasensitive Solution-Processed Photodetectors”, Adv. Mater. 2008, 20 (23), 4398–4402. [451] V. Sukhovatkin, S. Hinds, L. Brzozowski, E. H. Sargent, “Colloidal QuantumDot Photodetectors Exploiting Multiexciton Generation”, Science 2009, 324 (5934), 1542–1544. [452] J. P. Clifford, G. Konstantatos, K. W. Johnston, S. Hoogland, L. Levina, E. H. Sargent, “Fast, Sensitive and Spectrally Tuneable Colloidal-Quantum-Dot Photodetectors”, Nat. Nanotechnol. 2009, 4 (1), 40–44. [453] T. Rauch, M. B¨oberl, S. F. Tedde, J. Furst, M. V. Kovalenko, G. Hesser, U. Lemmer, W. Heiss, O. Hayden, “Near-Infrared Imaging with QuantumDot-Sensitized Organic Photodiodes”, Nat. Photon. 2009, 3 (6), 332–336. [454] B. N. Pal, I. Robel, A. Mohite, R. Laocharoensuk, D. J. Werder, V. I. Klimov, “High-Sensitivity p-n Junction Photodiodes Based on PbS Nanocrystal Quantum Dots”, Adv. Funct. Mater. 2012, 22 (8), 1741–1748. [455] M. B¨oberl, M. V. Kovalenko, G. Pillwein, G. Brunthaler, W. Heiss, “Quantum Dot Nanocolumn Photodetectors for Light Detection in the Infrared”, Appl. Phys. Lett. 2008, 92 (26), 261113. [456] H. Kim, K. Cho, H. Song, B. Min, J.-S. Lee, G.-T. Kim, S. Kim, S. H. Kim, T. Noh, “Photocurrent Mechanism in a Hybrid System of 1-ThioglycerolCapped HgTe Nanoparticles”, Appl. Phys. Lett. 2003, 83 (22), 4619–4621. [457] H. Kim, K. Cho, B. Park, J.-H. Kim, J. W. Lee, S. Kim, T. Noh, E. Jang, “Optoelectronic Characteristics of Close-Packed HgTe Nanoparticles in the Infrared Range”, Solid State Commun. 2006, 137 (6), 315–319. [458] M. Chen, H. Yu, S. V. Kershaw, H. Xu, S. Gupta, F. Hetsch, A. L. Rogach, N. Zhao, “Fast, Air-Stable Infrared Photodetectors Based on Spray-Deposited Aqueous HgTe Quantum Dots”, Adv. Funct. Mater. 2014, 24 (1), 53–59.

194

Bibliography [459] Y. Jin, J. Wang, B. Sun, J. C. Blakesley, N. C. Greenham, “SolutionProcessed Ultraviolet Photodetectors Based on Colloidal ZnO Nanoparticles”, Nano Lett. 2008, 8 (6), 1649–1653. [460] F. Guo, B. Yang, Y. Yuan, Z. Xiao, Q. Dong, Y. Bi, J. Huang, “A Nanocomposite Ultraviolet Photodetector based on Interfacial Trap-Controlled Charge Injection”, Nat. Nanotechnol. 2012, 7 (12), 798–802. [461] G. Konstantatos, L. Levina, J. Tang, E. H. Sargent, “Sensitive SolutionProcessed Bi2 S3 Nanocrystalline Photodetectors”, Nano Lett. 2008, 8 (11), 4002–4006. [462] T. P. Osedach, S. M. Geyer, J. C. Ho, A. C. Arango, M. G. Bawendi, V. Bulovi´c, “Lateral Heterojunction Photodetector Consisting of Molecular Organic and Colloidal Quantum Dot Thin Films”, Appl. Phys. Lett. 2009, 94 (4), 043307. [463] S. A. McDonald, P. W. Cyr, L. Levina, E. H. Sargent, “Photoconductivity from PbS-Nanocrystal/Semiconducting Polymer Composites for SolutionProcessible, Quantum-Size Tunableinfrared Photodetectors”, Appl. Phys. Lett. 2004, 85 (11), 2089–2091. ˇ c, “Efficient Polymer-Nanocrystal [464] D. Qi, M. Fischbein, M. Drndi´c, S. Selmi´ Quantum-Dot Photodetectors”, Appl. Phys. Lett. 2005, 86 (9), 093103. [465] K. R. Choudhury, Y. Sahoo, T. Y. Ohulchanskyy, P. N. Prasad, “Efficient Photoconductive Devices at Infrared Wavelengths using Quantum Dot-Polymer Nanocomposites”, Appl. Phys. Lett. 2005, 87 (7), 073110. [466] A. Biebersdorf, R. Dietm¨ uller, A. S. Susha, A. L. Rogach, S. K. Poznyak, D. V. Talapin, H. Weller, T. A. Klar, J. Feldmann, “Semiconductor Nanocrystals Photosensitize C60 Crystals”, Nano Lett. 2006, 6 (7), 1559–1563. [467] K. Szendrei, F. Cordella, M. V. Kovalenko, M. B¨oberl, G. Hesser, M. Yarema, D. Jarzab, O. V. Mikhnenko, A. Gocalinska, M. Saba, F. Quochi, A. Mura, G. Bongiovanni, P. W. M. Blom, W. Heiss, M. A. Loi, “Solution-Processable Near-IR Photodetectors Based on Electron Transfer from PbS Nanocrystals to Fullerene Derivatives”, Adv. Mater. 2009, 21 (6), 683–687. [468] T. P. Osedach, N. Zhao, S. M. Geyer, L.-Y. Chang, D. D. Wanger, A. C. Arango, M. C. Bawendi, V. Bulovi´c, “Interfacial Recombination for Fast Operation of a Planar Organic/QD Infrared Photodetector”, Adv. Mater. 2010, 22 (46), 5250–5254. [469] H. C. Shim, S. Jeong, C.-S. Han, “Controlled Assembly of CdSe/MWNT Hybrid Material and its Fast Photoresponse with Wavelength Selectivity”, Nanotechnology 2011, 22 (16), 165201.

195

Bibliography [470] L. Gao, D. Dong, J. He, K. Qiao, F. Cao, M. Li, H. Liu, Y. Cheng, J. Tang, H. Song, “Wearable and Sensitive Heart-Rate Detectors based on PbS Quantum Dot and Multiwalled Carbon Nanotube Blend Film”, Appl. Phys. Lett. 2014, 105 (15), 153702. [471] J. Zaumseil, H. Sirringhaus, “Electron and Ambipolar Transport in Organic Field-Effect Transistors”, Chem. Rev. 2007, 107 (4), 1296–1323. [472] S. Z. Bisri, C. Piliego, J. Gao, M. A. Loi, “Outlook and Emerging Semiconducting Materials for Ambipolar Transistors”, Adv. Mater. 2014, 26 (8), 1176–1199. [473] J. Tang, E. H. Sargent, Colloidal Quantum Dot Optoelectronics and Photovoltaics, Cambridge University Press, 2013. [474] M. S. Kang, C. D. Frisbie, “A Pedagogical Perspective on Ambipolar FETs”, ChemPhysChem 2013, 14 (8), 1547–1552. ¨ [475] S. Thiemann, S. J. Sachnov, F. Pettersson, R. Bollstr¨om, R. Osterbacka, P. Wasserscheid, J. Zaumseil, “Cellulose-Based Ionogels for Paper Electronics”, Adv. Funct. Mater. 2014, 24 (5), 625–634. [476] J. Zaumseil, X. Ho, J. R. Guest, G. P. Wiederrecht, J. A. Rogers, “Electroluminescence from Electrolyte-Gated Carbon Nanotube Field-Effect Transistors”, ACS Nano 2009, 3 (8), 2225–2234. [477] R. P. Ortiz, A. Facchetti, T. J. Marks, “High-k Organic, Inorganic, and Hybrid Dielectrics for Low-Voltage Organic Field-Effect Transistors”, Chem. Rev. 2010, 110 (1), 205–239. [478] S. Z. Bisri, C. Piliego, M. Yarema, W. Heiss, M. A. Loi, “Low Driving Voltage and High Mobility Ambipolar Field-Effect Transistors with PbS Colloidal Nanocrystals”, Adv. Mater. 2013, 25 (31), 4309–4314. [479] S. B. Grimm, F. Jakubka, S. P. Schießl, F. Gannott, J. Zaumseil, “Mapping Charge-Carrier Density Across the p-n Junction in Ambipolar CarbonNanotube Networks by Raman Microscopy”, Adv. Mater. 2014, 26 (47), 7986–7992. [480] M. Ha, J.-W. T. Seo, P. L. Prabhumirashi, W. Zhang, M. L. Geier, M. J. Renn, C. H. Kim, M. C. Hersam, C. D. Frisbie, “Aerosol Jet Printed, Low Voltage, Electrolyte Gated Carbon Nanotube Ring Oscillators with Sub-5 µs Stage Delays”, Nano Lett. 2013, 13 (3), 954–960. [481] K. Hong, S. H. Kim, K. H. Lee, C. D. Frisbie, “Printed, sub-2V ZnO Electrolyte Gated Transistors and Inverters on Plastic”, Adv. Mater. 2013, 25 (25), 3413–3418.

196

Bibliography [482] J. Lee, M. J. Panzer, Y. He, T. P. Lodge, C. D. Frisbie, “Ion Gel Gated Polymer Thin-Film Transistors”, J. Am. Chem. Soc. 2007, 129 (15), 4532– 4533. [483] R. Misra, M. McCarthy, A. F. Hebard, “Electric Field Gating with Ionic Liquids”, Appl. Phys. Lett. 2007, 90 (5), 052905. [484] K. Ghandi, “A Review of Ionic Liquids, Their Limits and Applications”, Green Sustainable Chem. 2014, 4 (1), 44–53. [485] D. Z. R. Feng, Y. Guo, “Revisiting Characteristics of Ionic Liquids: A Review for Further Application Development”, J. Environ. Prot. 2010, 1 (2), 95–104. [486] S. Thiemann, S. J. Sachnov, M. Gruber, F. Gannott, S. Spallek, M. Schweiger, J. Kr¨ uckel, J. Kaschta, E. Spiecker, P. Wasserscheid, J. Zaumseil, “SprayCoatable Ionogels Based on Silane-Ionic Liquids for Low Voltage, Flexible, Electrolyte-Gated Organic Transistors”, J. Mater. Chem. C 2014, 2 (13), 2423–2430. [487] J. Zaumseil, C. L. Donley, J.-S. Kim, R. H. Friend, H. Sirringhaus, “Efficient Top-Gate, Ambipolar, Light-Emitting Field-Effect Transistors Based on a Green-Light-Emitting Polyfluorene”, Adv. Mater. 2006, 18 (20), 2708–2712. [488] J. Zaumseil, R. H. Friend, H. Sirringhaus, “Spatial Control of the Recombination Zone in an Ambipolar Light-Emitting Organic Transistor”, Nat. Mater. 2006, 5 (1), 69–74. [489] A. Hepp, H. Heil, W. Weise, M. Ahles, R. Schmechel, H. von Seggern, “LightEmitting Field-Effect Transistor Based on a Tetracene Thin Film”, Phys. Rev. Lett. 2003, 91 (15), 157406. [490] M. Ahles, A. Hepp, R. Schmechel, H. von Seggern, “Light Emission from a Polymer Transistor”, Appl. Phys. Lett. 2004, 84 (3), 428–430. [491] J. S. Swensen, C. Soci, A. J. Heeger, “Light Emission from an Ambipolar Semiconducting Polymer Field-Effect Transistor”, Appl. Phys. Lett. 2005, 87 (25), 253511. [492] T. Takahashi, T. Takenobu, J. Takeya, Y. Iwasa, “Ambipolar Light-Emitting Transistors of a Tetracene Single Crystal”, Adv. Funct. Mater. 2007, 17 (10), 1623–1628. [493] Y. J. Zhang, T. Oka, R. Suzuki, J. T. Ye, Y. Iwasa, “Electrically Switchable Chiral Light-Emitting Transistor”, Science 2014, 344 (6185), 725–728. [494] R. S. Sundaram, M. Engel, A. Lombardo, R. Krupke, A. C. Ferrari, P. Avouris, M. Steiner, “Electroluminescence in Single Layer MoS2 ”, Nano Lett. 2013, 13 (4), 1416–1421.

197

Bibliography [495] M. Freitag, J. Chen, J. Tersoff, J. C. Tsang, Q. Fu, J. Liu, P. Avouris, “Mobile Ambipolar Domain in Carbon-Nanotube Infrared Emitters”, Phys. Rev. Lett. 2004, 93 (7), 076803. [496] F. Jakubka, C. Backes, F. Gannott, U. Mundloch, F. Hauke, A. Hirsch, J. Zaumseil, “Mapping Charge Transport by Electroluminescence in ChiralitySelected Carbon Nanotube Networks”, ACS Nano 2013, 7 (8), 7428–7435. [497] M. S. Kang, J. Lee, D. J. Norris, C. D. Frisbie, “High Carrier Densities Achieved at Low Voltages in Ambipolar PbSe Nanocrystal Thin-Film Transistors”, Nano Lett. 2009, 9 (11), 3848–3852. [498] J.-S. Lee, E. V. Shevchenko, D. V. Talapin, “Au-PbS Core-Shell Nanocrystals: Plasmonic Absorption Enhancement and Electrical Doping via Intra-particle Charge Transfer”, J. Am. Chem. Soc. 2008, 130 (30), 9673–9675. [499] H. Kim, K. Cho, D.-W. Kim, H.-R. Lee, S. Kim, “Bottom- and TopGate Field-Effect Thin-Film Transistors with p Channels of Sintered HgTe Nanocrystals”, Appl. Phys. Lett. 2006, 89 (17), 173107. [500] J. J. Urban, D. V. Talapin, E. V. Shevchenko, C. R. Kagan, C. B. Murray, “Synergism in Binary Nanocrystal Superlattices Leads to Enhanced p-Type Conductivity in Self-Assembled PbTe/Ag2 Te Thin Films”, Nat. Mater. 2007, 6 (2), 115–121. [501] M. V. Kovalenko, W. Heiss, E. V. Shevchenko, J.-S. Lee, H. Schwinghammer, A. P. Alivisatos, D. V. Talapin, “SnTe Nanocrystals: A New Example of Narrow-Gap Semiconductor Quantum Dots”, J. Am. Chem. Soc. 2007, 129 (37), 11354–11355. [502] D.-W. Kim, J. Jang, H. Kim, K. Cho, S. Kim, “Electrical Characteristics of HgTe Nanocrystal-Based Thin Film Transistors Fabricated on Flexible Plastic Substrates”, Thin Solid Films 2008, 516 (21), 7715–7719. [503] Y. Liu, M. Gibbs, C. L. Perkins, J. Tolentino, M. H. Zarghami, J. Bustamante, M. Law, “Robust, Functional Nanocrystal Solids by Infilling with Atomic Layer Deposition”, Nano Lett. 2011, 11 (12), 5349–5355. [504] D. K. Kim, A. T. Fafarman, B. T. Diroll, S. H. Chan, T. R. Gordon, C. B. Murray, C. R. Kagan, “Solution-Based Stoichiometric Control over Charge Transport in Nanocrystalline CdSe Devices”, ACS Nano 2013, 7 (10), 8760– 8770. [505] Y. Liu, J. Tolentino, M. Gibbs, R. Ihly, C. L. Perkins, Y. Liu, N. Crawford, J. C. Hemminger, M. Law, “PbSe Quantum Dot Field-Effect Transistors with Air-Stable Electron Mobilities above 7 cm2 V−1 s−1 ”, Nano Lett. 2013, 13 (4), 1578–1587.

198

Bibliography [506] S. J. Oh, N. E. Berry, J.-H. Choi, E. A. Gaulding, T. Paik, S.-H. Hong, C. B. Murray, C. R. Kagan, “Stoichiometric Control of Lead Chalcogenide Nanocrystal Solids to Enhance Their Electronic and Optoelectronic Device Performance”, ACS Nano 2013, 7 (3), 2413–2421. [507] B. Park, “Ligand-Free Ammonium Sulfide-Treated PbS Nanocrystal Field Effect Transistors”, Thin Solid Films 2014, 550, 696–699. [508] M. Scheele, D. Hanifi, D. Zherebetskyy, S. T. Chourou, S. Axnanda, B. J. Rancatore, K. Thorkelsson, T. Xu, Z. Liu, L.-W. Wang, Y. Liu, A. P. Alivisatos, “PbS Nanoparticles Capped with Tetrathiafulvalenetetracarboxylate: Utilizing Energy Level Alignment for Efficient Carrier Transport”, ACS Nano 2014, 8 (3), 2532–2540. [509] C. R. Bullen, P. Mulvaney, “Nucleation and Growth Kinetics of CdSe Nanocrystals in Octadecene”, Nano Lett. 2004, 4 (12), 2303–2307. [510] J. Schornbaum, B. Winter, S. P. Schießl, F. Gannott, G. Katsukis, D. M. Guldi, E. Spiecker, J. Zaumseil, “Epitaxial Growth of PbSe Quantum Dots on MoS2 Nanosheets and their Near-Infrared Photoresponse”, Adv. Funct. Mater. 2014, 24 (37), 5798–5806. [511] M. Hesse, H. Meier, B. Zeeh, Spektroskopische Methoden in der organischen Chemie, Bd. 7, Georg Thieme Verlag, Stuttgart, 2005. [512] J. C. de Mello, H. F. Wittmann, R. H. Friend, “An Improved Experimental Determination of External Photoluminescence Quantum Efficiency”, Adv. Mater. 1997, 9 (3), 230–232. [513] T.-S. Ahn, R. O. Al-Kaysi, A. M. M¨ uller, K. M. Wentz, C. J. Bardeen, “SelfAbsorption Correction for Solid-State Photoluminescence Quantum Yields Obtained from Integrating Sphere Measurements”, Rev. Sci. Instrum. 2007, 78 (8), 086105. [514] D. Ricard, M. Ghanassi, M. Schanne-Klein, “Dielectric Confinement and the Linear and Nonlinear Optical Properties of Semiconductor-Doped Glasses”, Opt. Commun. 1994, 108 (4-6), 311–318. [515] M. T. Trinh, A. J. Houtepen, J. M. Schins, T. Hanrath, J. Piris, W. Knulst, A. P. L. M. Goossens, L. D. A. Siebbeles, “In Spite of Recent Doubts Carrier Multiplication Does Occur in PbSe Nanocrystals”, Nano Lett. 2008, 8 (6), 1713–1718. [516] J. H. Warner, E. Thomsen, A. R. Watt, N. R. Heckenberg, H. RubinszteinDunlop, “Time-Resolved Photoluminescence Spectroscopy of Ligand-Capped PbS Nanocrystals”, Nanotechnology 2005, 16 (2), 175. [517] J. Schornbaum, Y. Zakharko, M. Held, S. Thiemann, F. Gannott, J. Zaumseil, “Light-Emitting Quantum Dot Transistors: Emission at High Charge Carrier Densities”, Nano Lett. 2015, 15 (3), 1822–1828.

199

Bibliography [518] J. F. Watts, J. Wolstenholme, An Introduction to Surface Analysis by XPS and AES, John Wiley & Sons Ltd, 2003. [519] R. F. Egerton, Physical Principles of Electron Microscopy, Springer Science+Business Media, Inc., 2005. [520] L. Reimer, H. Kohl, Transmission Electron Microscopy, 5. Aufl., Springer Science+Business Media, Inc., 2008. [521] D. K. Schroder, Semiconductor Material and Device Characterization, 3. Aufl., John Wiley & Sons, Inc., 2006. [522] Z. Saghi, P. A. Midgley, “Electron Tomography in the (S)TEM: From Nanoscale Morphological Analysis to 3D Atomic Imaging”, Annu. Rev. Mater. Res. 2012, 42 (1), 59–79. [523] J. Frank, Electron Tomography: Methods of Three-Dimensional Visualization of Structures in the Cell, 2. Aufl., Springer Science+Business Media, LLC, 2005. [524] P. Gilbert, “Iterative Methods for the Three-dimensional Reconstruction of an Object from Projections”, J. Theor. Biol. 1972, 36 (1), 105–117. [525] S. Thiemann, S. Sachnov, S. Porscha, P. Wasserscheid, J. Zaumseil, “Ionic Liquids for Electrolyte-Gating of ZnO Field-Effect Transistors”, J. Phys. Chem. C 2012, 116 (25), 13536–13544. [526] J.-F. Chang, M. C. Gwinner, M. Caironi, T. Sakanoue, H. Sirringhaus, “Conjugated-Polymer-Based Lateral Heterostructures Defined by HighResolution Photolithography”, Adv. Funct. Mater. 2010, 20 (17), 2825–2832. [527] J. Schornbaum, B. Winter, S. P. Schießl, B. Butz, E. Spiecker, J. Zaumseil, “Controlled In Situ PbSe Quantum Dot Growth around Single-Walled Carbon Nanotubes: A Noncovalent PbSe-SWNT Hybrid Structure”, Chem. Mater. 2013, 25 (13), 2663–2669. [528] A. J. Houtepen, R. Koole, D. Vanmaekelbergh, J. Meeldijk, S. G. Hickey, “The Hidden Role of Acetate in the PbSe Nanocrystal Synthesis”, J. Am. Chem. Soc. 2006, 128 (21), 6792–6793. [529] X. Chen, X. Huang, L. Kong, Z. Guo, X. Fu, M. Li, J. Liu, “Walnut-Like CdS Micro-Particles/Single-Walled Carbon Nanotube Hybrids: One-Step Hydrothermal Route to Synthesis and their Properties”, J. Mater. Chem. 2010, 20 (2), 352–359. [530] D. V. Talapin, E. V. Shevchenko, C. B. Murray, A. V. Titov, P. Kr´al, “DipoleDipole Interactions in Nanoparticle Superlattices”, Nano Lett. 2007, 7 (5), 1213–1219.

200

Bibliography [531] J. L. Blackburn, H. Chappell, J. M. Luther, A. J. Nozik, J. C. Johnson, “Correlation between Photooxidation and the Appearance of Raman Scattering Bands in Lead Chalcogenide Quantum Dots”, J. Phys. Chem. Lett. 2011, 2 (6), 599–603. [532] T. Li, G. Galli, “Electronic Properties of MoS2 Nanoparticles”, J. Phys. Chem. C 2007, 111 (44), 16192–16196. [533] E. Spiecker, “Interface and Surface Phenomena of Transition Metal Dichalcogenide Crystals: Ultrathin Layer Epitaxy and Self-Assembling of Nanostructures”, Microsc. Microanal. 2003, 9, 300–301. [534] S. Kuypers, J. V. Landuyt, S. Amelinckx, “Incommensurate Misfit Layer Compounds of the Type “MTS3 ” (M = Sn, Pb, Bi, Rare Earth Elements; T = Nb, Ta): A Study by Means of Electron Microscopy”, J. Solid State Chem. 1990, 86 (2), 212–232. [535] I. V. Lightcap, P. V. Kamat, “Fortification of CdSe Quantum Dots with Graphene Oxide. Excited State Interactions and Light Energy Conversion”, J. Am. Chem. Soc. 2012, 134 (16), 7109–7116. [536] H. Nam, S. Wi, H. Rokni, M. Chen, G. Priessnitz, W. Lu, X. Liang, “MoS2 Transistors Fabricated via Plasma-Assisted Nanoprinting of Few-Layer MoS2 Flakes into Large-Area Arrays”, ACS Nano 2013, 7 (7), 5870–5881. [537] Z. Yin, H. Li, H. Li, L. Jiang, Y. Shi, Y. Sun, G. Lu, Q. Zhang, X. Chen, H. Zhang, “Single-Layer MoS2 Phototransistors”, ACS Nano 2012, 6 (1), 74–80. [538] M. V. Artemyev, U. Woggon, H. Jaschinski, L. I. Gurinovich, S. V. Gaponenko, “Spectroscopic Study of Electronic States in an Ensemble of ClosePacked CdSe Nanocrystals”, J. Phys. Chem. B 2000, 104 (49), 11617–11621. [539] S. van Reenen, M. V. Vitorino, S. C. J. Meskers, R. A. J. Janssen, M. Kemerink, “Photoluminescence Quenching in Films of Conjugated Polymers by Electrochemical Doping”, Phys. Rev. B 2014, 89 (20), 205206. [540] K. V. Reich, T. Chen, A. L. Efros, B. I. Shklovskii, “Photoluminescence in Arrays of Doped Semiconductor Nanocrystals”, Phys. Rev. B 2013, 88 (24), 245311. [541] F. Jakubka, S. B. Grimm, Y. Zakharko, F. Gannott, J. Zaumseil, “Trion Electroluminescence from Semiconducting Carbon Nanotubes”, ACS Nano 2014, 8 (8), 8477–8486. [542] C. Galland, Y. Ghosh, A. Steinbruck, M. Sykora, J. A. Hollingsworth, V. I. Klimov, H. Htoon, “Two Types of Luminescence Blinking Revealed by Spectroelectrochemistry of Single Quantum Dots”, Nature 2011, 479 (7372), 203–207.

201

Bibliography [543] D. R. Lide, CRC Handbook of Chemistry and Physics, 90. Aufl., CRC Press, 2010. [544] J. T. Stewart, L. A. Padilha, M. M. Qazilbash, J. M. Pietryga, A. G. Midgett, J. M. Luther, M. C. Beard, A. J. Nozik, V. I. Klimov, “Comparison of Carrier Multiplication Yields in PbS and PbSe Nanocrystals: The Role of Competing Energy-Loss Processes”, Nano Lett. 2012, 12 (2), 622–628. [545] M. S. Kang, A. Sahu, D. J. Norris, C. D. Frisbie, “Size-Dependent Electrical Transport in CdSe Nanocrystal Thin Films”, Nano Lett. 2010, 10 (9), 3727– 3732. [546] S. Brovelli, C. Galland, R. Viswanatha, V. I. Klimov, “Tuning Radiative Recombination in Cu-Doped Nanocrystals via Electrochemical Control of Surface Trapping”, Nano Lett. 2012, 12 (8), 4372–4379. [547] J. Mooney, M. M. Krause, J. I. Saari, P. Kambhampati, “A Microscopic Picture of Surface Charge Trapping in Semiconductor Nanocrystals”, J. Chem. Phys 2013, 138 (20), 204705. [548] M. D. Peterson, L. C. Cass, R. D. Harris, K. Edme, K. Sung, E. A. Weiss, “The Role of Ligands in Determining the Exciton Relaxation Dynamics in Semiconductor Quantum Dots”, Annu. Rev. Phys. Chem. 2014, 65 (1), 317–339.

202

Appendix

A List of Abbreviations [EMIM][FAP]

1-ethyl-3-methyl-imidazolium tris(pentafluoroethyl)-trifluorophosphate

[EMIM][TFSI]

1-ethyl-3-methyl-imidazolium bis(trifluormethylsulfonyl) imide

ALD

atomic layer deposition

BDT

benzenedithiol

BF-TEM

bright field transmission electron microscopy

CB

conduction band

CBED

convergent beam electron diffraction

CCD

charge-coupled device

CNT

carbon nanotube

CoMoCAT

carbon nanotubes produced with a Co-Mo catalyst

CQD

colloidal quantum dot

CVD

chemical vapor deposition

DF-TEM

dark field transmission electron microscopy

DMF

dimethylformamide

DNA

deoxyribonucleic acid

DOS

density of states

EDL

electric double layer

EDT

ethandithiol

EDX

energy-dispersive X-ray spectroscopy

EL

electroluminescence

EQE

external quantum efficiency

FET

field-effect transistor

FIB

focused ion beam

FLG

few-layered graphene

FRET

F¨orster resonance energy transfer

FT

Fourier-transform

FWHM

full-width half maximum

HAADF

high-angle annular dark field

HiPCO

high-pressure carbon monoxide conversion

205

HOMO

highest occupied molecular orbital

HRTEM

high-resolution transmission electron microscopy

ICSD

inorganic crystal structure database

IR

infrared

IRF

instrument response function

LBL

layer-by-layer

LED

light-emitting diode

LEFET

light-emitting field-effect transistor

LUMO

lowest unoccupied molecular orbital

MEG

multi-exciton generation

MHA

6-mercaptohexanoic acid

MOA

8-mercaptooctanoic acid

MPA

3-mercaptopropionic acid

MWNT

multi-walled carbon nanotube

NA

numerical aperture

N EP

noise-equivalent power

NIR

near-infrared

NMP

n-methyl-2-pyrrolidone

oDCB

o-dichlorobenzene

P(VDF-HFP)

poly(vinylidene fluoride-cohexafluoropropylene)

P3HT

poly(3-hexylthiophene)

PCBM

[6,6]-phenyl-C61-butyric acid methyl ester

PDMS

polydimethylsiloxane

PET

polyethylene terephthalate

PI

polyimide

PL

photoluminescence

PMMA

poly(methyl methacrylate)

PTFE

polytetrafluoroethylene

QD

quantum dot

QY

quantum yield

RBM

radial breathing mode

rGO

reduced graphene oxide

SAED

selected-area electron diffraction

SDBS

sodium dodecylbenzene sulfonate

SDS

sodium dodecyl sulfate

SEM

scanning electron microscopy

206

STEM

scanning transmission electron microscopy

SWNT

single-walled carbon nanotube

TDMAH

tetrakis(dimethylamino)hafnium

TEM

transmission electron microscopy

TMD

transition metal dichalcogenide

TMS

bis(trimethylsilyl) sulfide

TOP

tri-n-octylphosphine

VB

valence band

XPS

X-ray photoelectron spectroscopy

207

B List of Symbols A

capacitor area

a

lattice constant

A0

intensity of decay curve

a1 ,a2

unit vectors of graphite

αB

Bohr radius

α

absorption coefficient

c

speed of light

Ch

chiral vector

Ci

capacitance

c(QD)

concentration of the quantum dots

D∗

normalized detectivity

d

diameter

dDL

thickness of dielectric layer

E

energy

EB

binding energy

Ef

electric field

EF

Fermi level

Eg

band gap

EK

kinetic energy

ε

relative permittivity

ε0

permittivity of vacuum

Et−

energetic level for electron traps

Et+

energetic level for hole traps

f

laser repetition rate

G

photoconductive gain

h

Planck’s constant

~

reduced Planck’s constant

Id

drain current

Id,ap

ambipolar drain current

Id,lin

linear drain current

Id,sat

saturation drain current

209

in

internal noise

IP h

photocurrent

k

wave vector

kP bS

extinction coefficient of bulk PbS

L

channel length

λ

wavelength

λAbs

wavelength of the absorption maximum

λExc

excitation wavelength

Lf ilm

thickness of quantum dot film

me

reduced mass of the electron

m∗e

effective mass of electrons

mh

reduced mass of the hole

m∗h

effective mass of holes

Mn

number average molecular weight

µ

charge carrier mobility

µrm

reduced mass

µe

electron mobility

µh

hole mobility

µlin

linear charge carrier mobility

µsat

saturation charge carrier mobility

MW

weight average molecular weight

N

number of absorbed photons per pulse

n

refractive index

Nep

emitted photons per second

Nph

photon flux per pulse

ν

frequency

P

output power

Pav

average laser power

Pin

incident optical power

Qmob

mobile charges

r

radius

rQD

radius of quantum dot

R

responsivity

rS

radius of diffraction limited focal spot

R(t)

weighted residual

S

sensitivity of a photodiode

210

Save

adapted sensitivity of a photodiode

σ

absorption cross section

τ

average lifetime

τc

carrier lifetime

τm

measured lifetime

τr

radiative lifetime

Tt

transit time

V

potential wall

Vbias

applied bias

vd

drift velocity

Vd

drain voltage

Vg

gate voltage

Vt

threshold voltage

W

channel width

WSp

work function of spectrometer

γnr

non-radiative decay rate

γnrX−

non-radiative decay rate of negative trions

γnrX+

non-radiative decay rate of positive trions

γnrX0

non-radiative decay rate

γrX−

radiative decay rate of negative trions

γrX+

radiative decay rate of positive trions

γrX0

radiative decay rate

211

C Publications Parts of this thesis have been published in the following articles: [1] J. Schornbaum, Y. Zakharko, M. Held, S. Thiemann, F. Gannott, J. Zaumseil “Light-emitting Quantum Dot Field-Effect-Transistors: Emission at High Charge Carrier Densities” Nano Lett. 2015, 15(3), 1822-1828. [2] J. Schornbaum, B. Winter, S. P. Schießl, F. Gannott, G. Katsukis, D. M. Guldi, E. Spiecker, J. Zaumseil “Epitaxial Growth of PbSe Quantum Dots on MoS2 Nanosheets and their Near-Infrared Photoresponse” Adv. Funct. Mater. 2014, 24(37), 5798-5806. [3] J. Schornbaum, B. Winter, S. P. Schießl, B. Butz, E. Spiecker, J. Zaumseil “Controlled In Situ PbSe Quantum Dot Growth around Single-Walled Carbon Nanotubes: A Noncovalent PbSe-SWNT Hybrid Structure” Chem. Mater. 2013, 25(13), 2663-2669.

213

D Acknowledgments Mein besonderer Dank gilt Frau Prof. Dr. Jana Zaumseil f¨ ur die Vergabe des interessanten Themas, die Freiheit alles ausprobieren und das Thema selbst formen zu d¨ urfen, das Vertrauen in meine Arbeit und die stets offene T¨ ur, um alle Probleme und Ideen zu diskutieren. Außerdem danke ich ihr, dass sie mir die Chance gegeben hat, meine Ergebnisse auf so vielen Konferenzen pr¨asentieren zu d¨ urfen! Vielen Dank auch an Prof. Dr. Alexander Eychm¨ uller und Prof. Dr. Erdmann Spiecker ¨ f¨ ur die Ubernahme des Zweit- und Drittgutachtens. Ich danke Benjamin Winter (Arbeitsgruppe Prof. Dr. Erdmann Spiecker) f¨ ur die Kooperation und Zusammenarbeit, ohne die es diese Arbeit so nicht geben w¨ urde. Er hat an meine Ideen geglaubt und mir dabei geholfen, viele der Resultate zu erhalten, die in dieser Arbeit pr¨asentiert werden. Vielen Dank f¨ ur die unz¨ahligen fachlichen und pers¨onlichen Gespr¨ache, alle Ermutigungen und Deine stets gute Laune! Many thanks to Dr. Yuriy Zakharko for contributing to many of the results. Sometimes it was hard to follow his confusing thoughts and theories, but in the end he managed to make it comprehensible for me. Many thanks for your continuous willingness to discuss and answer all my emails. Ich danke der ganzen NMOE Arbeitsgruppe (Dr. Florentina Gannott, Stefan Grimm, Martin Held, Dr. Florian Jakubka, Dr. Irina Lokteva, Katrin Ludwig, Marcel Rother, Stefan Schießl, Manuel Schweiger, Dr. Stefan Thiemann, Dr. Ming Wang, Dr. Yuriy Zakharko) f¨ ur die stets angenehme Arbeitsatmosph¨are. Sie haben es geschafft, dass ich mit Spaß in die Arbeit gekommen bin, auch wenn mich die Experimente mal frustrierten. Vielen Dank an meine beiden HiWis, Stefan Schießl und Felix Kalkowski, f¨ ur die Unterst¨ utzung bei der Synthese. Vielen Dank auch an die restlichen Kollegen des Lehrstuhls (OMD und LSP) f¨ ur die Hilfsbereitschaft und das nette Arbeitsklima und an Alfred Frey, Marco Heyder, Susanne Michler und Harald Rost f¨ ur die Unterst¨ utzung bei allen technischen und formellen Fragen. Many thanks to Prof. Ted Sargent (and the whole Sargent group) for the possibility to stay in his group for a few months and to learn a lot about QD processing. Thanks to the Graduate School in Advanced Optical Technologies and Prof. Dr. Jana Zaumseil for financing and supporting this research stay. Special thanks are due to Shokouh, Michael, Valerio, Ivan, and Silvia. They helped me to feel like home in Toronto and without them this research stay would not have been the same. Thanks a lot for the great time! Vielen Dank an Dr. Susanne Leubner und Dr. Stefanie Gabriel (beide Arbeitsgruppe Prof. Dr. Alexander Eychm¨ uller) f¨ ur den lehrreichen Synthesetag zu Beginn meiner Doktorarbeitszeit. Dieser eine Tag hat mich weiter gebracht als unz¨ahlige Synthesevorschriften zu lesen! Vielen Dank an Dr. Florentina Gannott und Anja Friedrich f¨ ur die REM Messungen, Helga Hildebrand f¨ ur die XPS Aufnahmen, Dr. Yuriy Zakharko f¨ ur die Hilfe bei der Fluoreszenzspektroskopie, Christel Dieker f¨ ur

215

die Pr¨aparation der FIB Proben, Benjamin Winter und Dr. Georgios Katsukis f¨ ur die Hilfe bei vielen TEM Messungen und Prof. Erdmann Spiecker f¨ ur die stetige Unterst¨ utzung bei allen TEM Messungen und deren Interpretationen! Diese Arbeit und die Konferenzbesuche wurden finanziell unterst¨ utzt vom Exzellenzcluster Engineering of Advanced Materials, dem DFG Projekt ZA 638/2 und den beiden Graduiertenschulen ”Graduate School Advanced Materials and Processes (GS AMP)” und der ”Graduate School in Advanced Optical Technologies (SAOT)”. F¨ ur das Korrekturlesen von Teilen dieser Arbeit und die vielen hilfreichen Tipps danke ich Benjamin Winter, Dr. Yuri Zakharko, Dr. Florian Jakubka, Dr. Georgios Katsukis und Michael Sekita! Vielen Dank an meine SAOT Kollegen Sasia, Felix und Michl f¨ ur all die sch¨one Zeit auf Seminaren, auf der Skipiste oder bei sonstigen Unternehmungen! Ein riesen Dank an Christian, Dieter und Michi - ohne euch w¨are das Studium nicht das gleiche gewesen! Es ist sch¨on, dass wir uns w¨ahrend der Promotion nicht aus den Augen verloren haben. Ich hoffe, wir werden noch das ein oder andere Kickermatch spielen! Nicht zuletzt ein großer Dank an Agnes und Alma - sie haben mich und meine Gedanken immer wieder ins normale Leben“ zur¨ uckgeholt. ” Ein besonderer Dank gilt Jonathan, der w¨ahrend der letzten Jahre oft auf gemeinsame Zeit verzichten musste und trotzdem immer Verst¨andnis f¨ ur mich hatte! Vielen lieben Dank an meine Familie, die mich immer meinen Weg hat gehen und meine Entscheidungen frei hat treffen lassen. Es ist sch¨on zu wissen, dass ich ein zu Hause habe, zu dem ich immer zur¨ uckkommen kann!

216

Smile Life

When life gives you a hundred reasons to cry, show life that you have a thousand reasons to smile

Get in touch

© Copyright 2015 - 2024 PDFFOX.COM - All rights reserved.