Metal-semiconductor hybrid nanoparticles: Halogen induced shape control, hybrid synthesis and electrical transport

©2014 Textbook 212 Pages


Metal-semiconductor hybrid nanoparticles combine materials with different physical properties in one nanostructure. Charge separation processes and potentially increased conductivity in thin film devices make them promising candidates for advanced applications in photocatalysis or (opto) electronics. The work on hand deals with the preparation of CdSe nanoparticles that later act as seeds for the defined deposition of metals and with the electrical characterisation of monolayers of the resulting hybrid structures. In context with the shape control of the semiconductor component in a hot injection synthesis, the role of halogen compounds and the influence of their molecular structure are examined. Analytically as well as theoretically supported explanations for the formation of the evolving hexagonal pyramidal shape, which is especially favourable for metal deposition, are presented. The deposition of different metals onto the obtained semiconductor components is examined and unusual instabilities of an Au shell on CdSe hybrid nanoparticles are investigated. Furthermore, the impact of deposited Pt on the electrical transport of CdSe nanopyramids is demonstrated.


Table Of Contents

7 Zusammenfassung
A Additional data
B Safety
Curriculum Vitae
Publications and conference contributions

List of Abbreviations
Binding energy
Conduction band
Density functional theory
n-Dodecyltrimethylammonium bromide
n-Dodecyltrimethylammonium chloride
Energy dispersive X-ray spectroscopy
Full width at half maximum
Highly oriented pyrolitic graphite
Inductively Coupled Plasma Optical Emission Spectroscopy
Metal induced gap state
Nearest-neighbour hopping
Oleic acid
Quantum yield

List of Abbreviations
Scanning transmission electron microscopy
Tetra-n-butylammonium borohydride
Transmission electron microscopy
Tri-n-octylphosphine (oxide)
Total reection X-ray uorescence spectroscopy
Valence band
Variable-range hopping
X-ray Photoelectron Spectroscopy
Powder X-ray diractommetry

List of Figures
2.1 Wurtzite crystal structure of CdSe . . . . . . . . . . . . . . . . . . . . . . .
2.2 Types of ligand coordination in nanoparticles . . . . . . . . . . . . . . . .
2.3 Factors inuencing the size and shape of nanoparticles . . . . . . . . . . .
2.4 LaMer plot with separated growth stages . . . . . . . . . . . . . . . . . . .
2.5 Evolution of CdSe absorption and morphology with the additive DCE . . .
2.6 Correlation of nanopyramid size and DCE/Cd ratio . . . . . . . . . . . . .
2.7 Micrographs of CdSe nanopyramids with oleic instead of phosphonic acid .
2.8 Temporal evolution of absorption maxima with dierent chloroalkanes . . .
2.9 TEM micrographs of samples prepared with DCE, DBE and DIE . . . . .
2.10 Addition of 1-chlorooctadecane after the nucleation . . . . . . . . . . . . .
2.11 XPS spectra of nanorods and -pyramids . . . . . . . . . . . . . . . . . . .
2.12 Temporal changes of the elemental composition determined by TXRF . . .
2.13 Rod and hexagonal pyramidal geometry with distinct facets . . . . . . . .
2.14 DFT simulations: PPA on dierent crystal facets . . . . . . . . . . . . . .
2.15 LaMer-type plot of the CdSe nanopyramid formation reaction . . . . . . .
2.16 Plot of the pH of aliquots versus time . . . . . . . . . . . . . . . . . . . . .
3.1 Band position and work functions of CdSe and dierent metals . . . . . . .
3.2 Modes of epitaxial heterodeposition. . . . . . . . . . . . . . . . . . . . . . .
3.3 Mechanisms of oligomer formation . . . . . . . . . . . . . . . . . . . . . . .
3.4 Scheme of reactive sites of CdSe nanopyramids. . . . . . . . . . . . . . . .
3.5 Micrographs of nanopyramids with cluster sized Au domains . . . . . . . .
3.6 Au domain growth and absorption spectra with increasing Au/CdSe ratio .
3.7 Photographs of Au precursor solutions . . . . . . . . . . . . . . . . . . . .
3.8 HR-TEM micrographs of CdSe nanopyramids with Au shell and dots. . . .
3.9 IR-spectra of CdSe nanopyramids before and after treatment with dode-
canethiol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

List of Figures
3.10 X-ray diraction pattern of Au-CdSe shell nanoparticles . . . . . . . . . .
3.11 Atomic ratios and diameters of Au-CdSe dot and shell samples . . . . . . .
3.12 XPS survey spectra of CdSe and Au-CdSe nanoparticles . . . . . . . . . .
3.13 XPS data of Se 3d and Au 4f regions of Au-CdSe samples with shell and dots. 67
3.14 TEM micrographs of annealed Au-CdSe lms . . . . . . . . . . . . . . . .
3.15TEM micrograph of nanopyramids after ion exchange with Ag. . . . . . . .
3.16 Diraction pattern after ion exchange from Cd to Ag . . . . . . . . . . . .
3.17 TEM micrographs of CdSe nanopyramids at dierent stages of cation ex-
change with Pd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.18 High resolution micrographs, EDX mapping and electron diraction of CdSe-
nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.19 Evolution of absorbance spectra during an ion exchange reaction between
Cd and Pd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.20 Oligomeric Pt-CdSe nanoparticles with dierent Pt domain sizes . . . . . .
3.21 Absorbance and emission of Pt-CdSe samples . . . . . . . . . . . . . . . .
3.22 Electron diraction and interface regions of Pt-CdSe nanoparticles . . . . .
3.23 EDX map of Pt-CdSe hybrid nanoparticles. . . . . . . . . . . . . . . . . .
3.24 TEM micrographs of in situ annealed Pt-CdSe nanoparticles . . . . . . . .
3.25High resolution micrographs of annealed Pt-CdSe nanoparticles . . . . . .
4.1 Scheme of a nanoparticle based source-drain device . . . . . . . . . . . . .
4.2 Coulomb blockade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3 Electron transport mechanisms in disordered semiconductors . . . . . . . .
4.4 Band structures and charge transport at metal-semiconductor junctions . .
4.5Interdigitated array electrodes . . . . . . . . . . . . . . . . . . . . . . . . .
4.6 Micrographs of a CdSe nanopyramid array . . . . . . . . . . . . . . . . . .
4.7 Current-voltage curves of a CdSe nanopyramid device . . . . . . . . . . . . 100
4.8 Micrographs of Pt-CdSe nanoparticles and devices . . . . . . . . . . . . . . 102
4.9 Dark and photocurrent of Pt-CdSe arrays (Pt= 1.7 nm) . . . . . . . . . . . 103
4.10 Dark and photocurrent of Pt-CdSe arrays (Pt= 3.2 nm) . . . . . . . . . . . 105
4.11 Fits of current-voltage curves of Pt-CdSe arrays . . . . . . . . . . . . . . . 106
4.12 Temperature dependence of dark currents (Pt= 3.2 nm) . . . . . . . . . . . 109
4.13 Possible paths for electron transfer in Pt-CdSe arrays . . . . . . . . . . . . 110
5.1 Lengths in hexagonal (di)pyramids . . . . . . . . . . . . . . . . . . . . . . 127

A.1 Temporal evolution of absorbance and emission in a reaction with DCE . .
A.2 Morphological evolution without additive . . . . . . . . . . . . . . . . . . .
A.3 Optical and morphological evolution in reactions with chloroalkanes . . . .
A.4 Electron beam induced migration in Au-CdSe samples . . . . . . . . . . . III
A.5 XPS Se 3d signal of a Au-CdSe shell sample . . . . . . . . . . . . . . . . . III
A.6 Beam current dependency of migration in Au-CdSe samples . . . . . . . . IV
A.7 Composition of nanoparticles after ion exchange with Ag . . . . . . . . . . IV
A.8 Evolution of absorbance spectra of CdSe nanopyramids incubated with Pd
at room temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A.9 X-ray diraction pattern of Pt-CdSe nanoparticles . . . . . . . . . . . . . .
A.10 Electrode with CdSe nanopyramid array after annealing. . . . . . . . . . . VI
A.11 Schematic depiction of a 4-point probe measurement . . . . . . . . . . . . VI
B.1 GHS-pictograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XXI
List of Schemes
3.1 Au deposition onto CdSe nanoparticles in the presence of amine ligands . .
List of Tables
2.1 Nanoparticle dimensions at dierent DCE/Cd ratios. . . . . . . . . . . . .
2.2 Adsorption energies of ligands on dominant CdSe surfaces . . . . . . . . .
3.1 Standard reduction potentials of relevant species . . . . . . . . . . . . . . .

List of Tables
3.2 Ionic radii, acid hardness, relevant selenides and their crystal structure(s)
of dierent metal cations . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3 Composition of nanopyramids before and after reaction with Pd(II) . . . .
3.4 Atomic composition of CdSe and CdSe-Pt samples determined by EDX. . .
4.1 Temperature dependencies of selected electrical transport mechanisms . . . 108
5.1 Details of reactions with dierent halogenated additives . . . . . . . . . . . 115
5.2 Details of Au deposition reactions . . . . . . . . . . . . . . . . . . . . . . . 120
5.3 Parameters for incubation of CdSe nanopyramids (DCE) with Pd(II) . . . 121
5.4 Parameters for incubation of CdSe nanopyramids (COD) with Pd(II) . . . 122
5.5 Parameters for incubation of CdSe nanopyramids (DCE) with Pd(II) 2 . . 122
B.1 Safety information for employed substances . . . . . . . . . . . . . . . . . . VII
B.2 All H, EUH and Pstatements . . . . . . . . . . . . . . . . . . . . . . . . . XI

1 Introduction
Based on their outstanding properties, nanoparticles have entered broad areas of research
related to (photo)catalysis [1], energy conversion [2], optoelectronics [3], and biomedicine
[4, 5].A major factor determining these properties and the reactivity of nanoparticles
is their size and related to it their large surface-to-volume ratio.The small dimensions
cause quantum mechanical connement of electrons and thus altered physical conditions
compared to bulk materials [6].The reduction in size is also advantageous to save precious
materials such as catalytic metals.With proceeding climate change and pollution, high
hopes rest on developments in solar energy conversion.Nanoparticles oer a variety of
solutions for related applications in photovoltaics and photocatalytic conversion of solar
into chemical energy.
During the past decades, the preparation of nanostructures consisting of single and
multiple components has developed into a tool box for the creation of purpose-designed
materials.Especially the colloidal synthesis can be applied to prepare nanoparticles in a
large variety of shapes and sizes with comparatively low eort and costs.With the intro-
duction of high temperature preparation methods twenty years ago [7], a way to obtain
highly monodisperse nanoparticles was paved.From then on, the control over size and
shape of the particles has grown steadily.
Another strong impulse was the selective formation of multicomponent nanostructures
that combine materials with dierent physical properties [8].In secondary steps other
materials can be grown onto prepared nanoparticles with high precision [9, 10].Metal-
lic nanostructures deposited onto semiconductors, for instance, facilitate the separation
of charges that are photogenerated in the semiconductor [11].This fundamental process
is the basis for improved eciencies in elds such as photocatalytic water splitting [12].
An interesting feature of metal domains on semiconductor nanoparticles is their ability to
improve electrical contacts to the latter [13].This might be utilised to increase the charge
transport in semiconductor nanoparticle arrays and aect the photocurrent obtained under

Creating and characterising such a nanoparticle array requires a high degree of control
over all steps involved, from semiconductor synthesis over metal deposition to the nal
assembly. To reach this control, an understanding of the fundamental processes accompa-
nying each step is necessary. By modulating the shape of the semiconductor material, for
example, the number of sites reactive towards the deposition of metals can be varied which
aects the whole architecture. Their comparatively large surface results in lowered melt-
ing points of nanoparticles and furthermore makes them susceptible for fast dissolution,
the adsorption of molecules and reactions with the surrounding medium [14]. To prevent
them from coagulation, nanoparticles are coated by a layer of surfactants, also referred
to as ligands or stabilisers. Apart from exhibiting a stabilising eect, the adsorption of
such ligands may play an important role in the shape evolution of wet-chemically prepared
nanostructures. In connection with the preparation of CdSe-carbon nanotube composites
1,2-dichloroethane was observed to induce the formation of hexagonal pyramidally shaped
CdSe nanoparticles [15, 16]. A mechanism involving chloride ions was presumed.
The peculiar pyramidal geometry provides a high number of sites prone to metal de-
position and is thus an ideal candidate for an envisaged synthesis of hybrid nanoparticles
with several dened metal domains in an oligomeric structure. The control of the shape
evolution in reactions without carbon nanotubes and a possible adaptation of the method
to develop dierent morphologies is desired. For this reason, a better understanding of
the inuence of the di-halogen alkane on the shape evolution shall be gained in this work.
Another goal is the preparation of hybrid nanoparticles with oligomeric structure for the
electrical characterisation in two-dimensional arrays. Finally, the obtained material shall
be assembled and pioneering electrical studies shall be conducted to nd out how the
interplay of the domains inuences two-dimensional conductance and the generation of
Following these steps, this thesis is separated into three chapters, each with a theoretical
introduction, a results and discussion as well as a conclusions section. In the concluding
sections, specic aspects concerning the results of the corresponding chapter will be treated.
In chapter 2, the interactions between ligands and the nanoparticle surface play an im-
portant role. The shape control of the semiconductor component by halogenated additives
is examined. To better understand why 1,2-dichloroethane induces such a peculiar shape
and if the eect may be exploited to tune the size and shape of nanoparticles, experimental
and theoretical studies are combined. Systematic variations of halogenated additives as
well as elemental and surface analysis are applied. In the theoretical part, calculations

based on the density functional theory and parallels to a classical crystal growth model are
In chapter 3, the seeded-growth deposition of four metals, Au, Ag, Pd and Pt, onto
hexagonal pyramidal CdSe nanoparticles in organic solution is examined. The reasons for
the formation of a presumed shell-like Au structure with undened composition observed
in preliminary work([17]) are scrutinised. Among the conducted experiments are studies
concerning the inuence of the oxidation state of the precursor on the morphology of
forming hybrids. The oxidation state of the metal in the shell is examined by X-ray photo-
electron spectroscopy. The deposition behaviour of the other three metals is tested by a
new common synthetic method with oleylamine as ligand and reducing agent for the metal.
Chapter 4 deals with electrical properties of nanoparticles and specically two-dimen-
sional Pt-CdSe hybrid nanoparticle arrays. Nanoparticles with two sizes of Pt domains
are examined. The assemblies are prepared with the Langmuir-Blodgett technique and
measured in a probe station under vacuum. Current-voltage curves are recorded in darkness
and under illumination conditions with dierent irradiation sources. Additionally, the
temperature is varied down to cryogenic conditions. In comparison to literature data of
related systems observations concerning the transport mechanisms are made.
Experimental details of all chapters are summarised in a joined experimental chapter
(chapter 5). A general summary with outlookfor the workfollows in chapter 6, while
chapter 7 contains a summary in German. Additional data for chapters 2, 3 and 4 is
provided in the appendix.

2 Halogen induced shape control of
CdSe nanoparticles
Due to their special size dependent optical and electronic properties semiconductor nanopar-
ticles nd application in photovoltaics [18, 19], (photo)catalysis [1, 12], light emitting de-
vices [20] and biological labelling [21, 22], among others [23].
In these contexts, the shape of the nanoparticles may become important due to physical
properties depending on the dimensionality of quantum connement in the nanostructure
or simply their packing behaviour [3, 24, 25, 26]. Apart from their morphology, an aspect
that is critical for applications is the passivation of the nanoparticle surface by stabilisers.
Owing to the successful implementation of nanoparticles partially capped by or post
synthetically treated with halides into solar cells with increased eciency [27, 28], incorpo-
ration of atomic halogen ligands has attracted much attention recently [29, 30, 31, 32, 33].
In addition to enhancing physical properties in nanoparticle arrays, halides show inter-
esting eects on the shaping of nanoparticles. With metal nanoparticles they are delib-
erately employed to control the growth with inuences on the shape formation [34, 35].
For semiconductor nanoparticles, several cases of wurtzite structures with hexagonal bul-
let, pyramid, pencil or diamond shape were reported which had in common that chloride
precursors were present [36, 37, 38]. A few studies showed an increase of morphological
uniformity in branched cadmium chalcogenides with wurtzite arms growing on seeds of
deviating crystal structures when halides were added [39, 38, 40]. Juárez and co-workers
observed how traces of 1,2-dichloroethane employed as solvent for carbon nanotubes added
in situ to a synthesis of CdSe nanorods induced a shape evolution [15, 16]. Hexagonal pyra-
midal nanoparticles with wurtzite structure evolved, which then formed composites with
the nanotubes. Chloroalkanes were furthermore reported to aid the preparation of sheet-
like lead sulde nanostructures with the chemical structure of the molecules inuencing the
dimensions of the crystals [41, 42]. These circumstances built a promising foundation for
further studies on the shape control of semiconductor nanoparticles by halogen compounds

with CdSe as a model system with well-known properties. An extension of feasible mor-
phologies presents an attractive goal with respect to increasing demands on the control of
nanoparticle shapes for application in thin lm arrays or as seed material with controllable
reactive sites for heteronanoparticle formation.
In the following, the most important theoretical aspects of the wet-chemical nanoparticle
synthesis and methods of semiconductor nanoparticle shape control will be introduced.
They will lay the basis for the discussion of aspects of the shape control of CdSe nanopar-
ticles aided by (organo) halogen compounds examined in this work. For more detailed
insights into properties and preparation of nanoparticles, references [23, 43, 44, 45, 46]
are suggested. A substantial part of the chapter is based on and reproduced in part
with permission from [Meyns, M., Iacono, F., Palencia, C., Geweke, J., Coderch, M. D.,
Fittschen, U. E. A., Gallego, J. M., Otero, R., Juárez, B. H., Klinke, C. Chem. Mater.
2014, 26, 1813-1821.] Copyright [2014] American Chemical Society.
2.1 Properties and synthesis of semiconductor
The eld of semiconductor nanoparticle synthesis oers a wide and growing variety of
synthetic protocols and factors that can be tuned to inuence and design the dimensions
and shapes of the crystals. Known morphologies range from zero- to two-dimensional
in terms of bulk-like dimensions (quantum dots [47], nanotubes and nanowires [48, 49],
tetrapods [50], nanosheets [51, 52]). This is appealing in so far as the shape of nanoparticles
inuences physical and chemical properties through changes in electric elds and crystal
facet dependent surface reactivity [14, 53, 54, 55].
2.1.1 Properties of semiconductor (CdSe) nanoparticles
In terms of electrical properties, semiconductors take a middle position between highly
conducting metals and insulating materials such as glasses or polymers. The reason for
this is that they are able to conduct electricity only when activated by thermal energy
or light. This additional energy allows electrons to move from the lower valence to the
upper conduction band across the band gap; in metals there is no gap between the bands,
whereas insulators are dened as having a band gap bigger than 4 eV. These bands, formed
by energetically close lying electron orbitals, are delocalised over the whole crystal and

PFI roperties nd synthesis of semiondutor nnoprtiles
exited eletronEhole pirs @exitonsA re free to move in ll diretions in ulk mterilsF
sn nnoprtilesD the numer of tomsD eh ontriuting one oritl to the ndD is muh
smller @etween IHH nd few IHHHHAD whih uses n inrese of the gp with deresing
nnoprtile size nd the formtion of disrete sttes t the nd edges TD PRF et sizes of
fewnnometresD the wve funtions of eletron nd hole eome on(ned in the dimensions
of the semiondutor nnoprtilesD nowlled quntum dotsD nd the gp is strongly size
dependent @quntum size e'etAF he numer of nnoprtile dimensions rnging in the
on(nement regime determines the density of ville energeti sttesF his wyD the
shpe nd direted growth of nnostruture my in)uene its nd gp STF sn spheril
nnoprtiles ll three dimensions re 'eted nd the eletroni levels develop into disrete
sttesF his is visile y lue shifts in sorption nd emission spetrD in whih the (rst
mximum t lower wvelengths in the sorption spetr n e employed to diretly
relte size nd nd gp of the nnoprtiles SUD SVF hue to the high surfeEtoEvolume
rtioD the on(gurtion of the outer lyer of toms nd the lignd sphere proteting the
nnoprtiles from ogultion mke lrge impt on the overll propertiesF por exmpleD
the quntum yield of emission nd the eletril trnsport in nnoprtile rrys n e
signi(ntly deteriorted y hrge trpping sites in form of dngling ondsF
he medium vlued diret nd gp of ssEs semiondutor gde @IFUR e in ulk SWA
mkes it one of the fewmterils whose nnoprtiles n e tuned in size to sor nd
emit light throughout the visile spetrumF he eletril properties of gde nnoprtiles
will e ddressed in ghpter RF gde rystls our in the ui zin lende or in the
hexgonl wurtzite strutureF he ltter is depited in pigure PFIF et room temperture
the ui phse is slightly more stle y IFR me per tom TIF olytypism ours
t higher tempertures nd the rystl struture depends on the retion onditionsF sn
the presene of hlogens preferentil formtion of the wurtzite phse ws oservedD the
reson for whih hs not een lri(ed so fr RHD TPD TQF he wurtzite struture is
nisotropiD whih eomes pprent y di'erenes in the numer of onds per element
when ounting in opposite diretions of the cExisF sn the positive diretion gd hs one
ond pointing upwrds wheres e hs threeF por this resonD hexgonl gde rystls re
usully terminted y positively hrged gdErih @HHHIA nd negtively hrged eErih
@HHH©IA fet to minimise the numer of dngling onds TRD TSD TTF his nisotropi hrge
distriution retes dipole moment in the nnoprtiles TUD TVF

grei PF revyqix sxhgih rei gyxyv yp ghi xexyesgvi
Figure 2.1: rexgonl wurtzite rystl struture of gde omposed from unit ells with the
lttie onstnts aD bX RFQH ÅD cX UFHI Å THY gd toms re eigeD e toms re
orngeF he upper @HHHIA lyer onsists of gdEtoms with one dngling ondD the
lower @HHH©IA fet is usully terminted y e with one dngling ondF greted with
Materials studioF
CdSe nanoparticles alone or as cores coated with inorganic or organic shell materials,
are applied in solar cells [18], light emitting diodes [20], and biological imaging [4]. Fur-
thermore, theyhave acted as model systems for the exploration of properties in a varietyof
contexts. Among these are optical and electrical studies as well as investigations of shape
control during synthesis [69, 70, 71].
2.1.2 Colloidal semiconductor nanoparticle synthesis and shape
For further processing and ecient applications nanoparticles should be homogeneous in
their size and shape. The method of choice to obtain nanoparticles with the highest
precision in terms of size and shape is the wet-chemical bottom-up approach, where crystals
are prepared from molecular precursors of the components. Semiconductor nanoparticles
with small size distributions of below 10% are mainlyobtained byhot-injection syntheses
with separated nucleation and growth stages as pioneered byMurray, Norris and Bawendi
in 1993 [7]. Before going into detail, the basics of nanoparticle formation and dierent
models of shape control will be presented briey.

2.1 Properties and synthesis of semiconductor nanoparticles Basics of nanoparticle nucleation
For the formation of nanoparticles starting materials are mixed and react to a solute form
of the crystal
, not yet considered as a solid phase, which evolves into nuclei once a critical
supersaturation is reached. The activation energy necessary for nucleation is provided
when the change of the overall free energy
G = -
m,N P
T ln(S) + 4r
reaches a maximum. In Equation (2.1) V
m,N P
is the molecular volume of the material in
the crystal, r is the radius of the spherical nuclei, k
is the Boltzmann constant, S is the
saturation ratio and is the surface free energy per unit surface area [14]. The maximum
is reached for a critical radius r
when the saturation ratio S =
with the solute
concentration c
and the solubility product of the crystal material k
is larger than
one [72]. The critical radius is obtained as
m,N P
T ln(S)
by derivation of Equation 2.1. During nucleation, the supersaturation depletes until no
further nuclei form. These nuclei continue to grow towards nanoparticles whose size and
shape can be controlled by growth parameters such as concentration, temperature and the
choice of ligands. Finally, the solute concentration will approach the level of the material's
solubility and the critical radius will shift to larger values so that smaller particles dissolve
in favour of further growth of bigger ones, a process known as Ostwald ripening [73]. Mechanisms of shape control
Non-spherical shapes form whenever the facets of a crystal exhibit dierent growth rates;
fast growing facets are eliminated and slowly growing facets eventually determine the shape
or Tracht, the entirety of the facets that form the crystal surface. Growth rates can be gov-
erned by thermodynamic or kinetic inuences, depending on the growth conditions. Apart
from temperature and concentration, a major role in size and shape control is played by
In literature the terms solute and monomer are employed to describe a small unit consisting of at least
one of each crystal components. Solute is the term found in works on classical crystal growth and is
less limited, since the solute form must not necessarily be a single formula unit.

organic surfactants, which are employed to stabilize the solid nanoparticle phase against
dissolution and coagulation [74, 75, 76]. The most commonly employed types of ligands
to stabilise semiconductor nanoparticles are amphiphilic organic molecules counted to the
groups of neutral L-type or negatively charged X-type ligands. The rst group may co-
ordinate both cationic and anionic crystal components, while the second one selectively
balances positively charged surface atoms to reach overall charge neutrality. If both types
are present, the interaction with X-type ligands will dominate due to stronger interactions
with the surface [29, 31, 77, 78]. L-type ligands provide additional stability but are more
easily washed away during purication. A breakthrough observation regarding selective
X-type adsorption is the case of phosphonic acid impurities in a CdSe synthesis with oth-
erwise neutral tri-n-octylphosphane (TOP) ligands in tri-n-octylphosphane oxide (TOPO),
which coordinated the nanoparticles in form of phosphonates [79]. As mentioned earlier,
halide ions acting as atomic X-type ligands gain increasing popularity.
The terms surfactant, stabiliser and ligand are often used interchangeably in the lit-
erature. Since surfactants and stabilising molecules usually are amphiphilic compounds
with alkyl chains, the term ligand will be applied here within the meaning of a potentially
surface bound constituent including short molecules and halides.
Figure 2.2: Most of the ligands employed in semiconductor nanoparticle stabilisation belong to
the groups of neutral L-type ligands, which bind to cations and anions of the crystal
(a dative bond to the metal cation is shown here), and X-type ligands that selectively
bind to the cationic component. The named examples are tri-n-alkylphosphanes
P), primary amines (RNH
), thiolates (RS
) and phosphonates ((RPO
Halides can act as atomic X-type ligands. In metallic nanoparticles with negative
surface charges cationic ligands are another important group of surfactants.

2.1 Properties and synthesis of semiconductor nanoparticles
Kumar and Nann summarised theories explaining the shape evolution of nanoparticles
and named ve important ones: (i) thermodynamic theory, (ii) selective-adsorption model,
(iii) oriented attachment model, (iv) molecular template theory and (v) eective-monomer
model [76].
(i) Thermodynamic theory
According to the thermodynamic theory, which is based on the works of Gibbs, Curie and
Wul, the shape of a crystal in equilibrium with its environment is determined by crystallo-
graphic facets whose combined surface free energy minimises the total free energy of the
system [80]. In a supersaturated solution equilibrium shapes with low index facets would
form because the growth of higher index facets is accelerated. In classical crystal growth
theory, thermodynamic control is one approach to understand the Habitus
of polyhedral
equilibrium forms of crystals [81]. An experimentally veried model to predict such forms
was developed by Wol and co-workers [66]. It focuses on the contribution of dangling
bonds to the surface free energy of a facet and thus includes their coordination and their
ionicity in form of the ratio between cationic and anionic bond energies into the calculation.
The traditional thermodynamic theory often does not account for shape evolutions in o-
equilibrium systems applied in nanoparticle synthesis. Furthermore, the Wol model has
not been related with modern nanoparticle synthesis and shapes so far, to the best of the
author's knowledge.
(ii) Selective-adsorption model
The selective-adsorption model was among the rst to explain anisotropic growth of II-VI
semiconductor nanoparticles such as CdSe. It is based on the assumption that ligands
preferably adsorb to certain crystallographic facets. With a mixture of ligands the dif-
ferent anities thus inuence the relative growth rates of evolving facets, which leads to
anisotropic shapes [75, 82].
(iii) Oriented attachment model
A reduction of surface energy can be evoked by attachment of already formed nanoparticles
via crystallographically similar facets. This process can be explained by thermodynamic
driving forces or dipole-dipole interactions [83, 84]. Such oriented attachment can lead to
The term Habitus should not be mixed up with Tracht. A Tracht is the set of crystal facets terminating
the surface of a crystal, the Habitus is determined by the relative size of these facets.

(iv) Molecular template theory
(v) Effective-monomer model The hot injection synthesis and shape control
f oc.

tions develop erly due to ystwld ripeningF ytherwiseD ripening sets in during the seond
stge of growth @defousingD stge sAF
he evolution of nnoprtile shpes in this type of synthesis is lso relted to the di'erE
ent retion stgesF ith high supersturtion in stge sss the ondition for kinetilly
ontrolled growth fter the e'etiveEmonomer model is ful(lled nd monodisperse dotE
nd rodEshped or tetrpodl nnoprtiles form UPF e high soluteGmonomer onentrE
tion nd rod growth of gdeD for exmpleD were otined y employing phosphoni ids
VWD WHD WPD whih formstle omplexes with gd in the preursor nd on the nnoprtile
surfe UUD WTD WUF rereD the e'etiveEmonomer model nd seletive dhesion of lignds
overlpD sine the eEterminted @HHH©IA fet in gde is the one lest pped y lignds
in the orresponding syntheti onditionsD so tht nnorods growing preferentilly in the
HHH©I diretion long the cExis re formed WVF sf the supersturtion is not mintined
t high level y dditionl preursor injetion or suitle preursor onversion kinetis
USD UWD WWD the depletion of preursor nd solute will led to ripening of the prtiles
whih results in the formtion of thermodynmilly more stle shpesF eng et al. reE
ported supersturtion dependent trnsition fromnisotropi @IhA to threeEdimensionl
growth followed y intrprtile ripening WHF xnorods reErrnged to spheril shpes
in limited di'usion sphere round them without mss exhnge with the ulk solutionF
ritions of the monomer onentrtion in the growth solution my led to di'erent
morphologies from dots to tetrpods under otherwise identil retion onditions UIF
he understnding of this type of retion is rod ut due to the omplexity of interE
tions etween di'erent omponents nd the sometimes lrge in)uene of impuritiesD for
exmple induing rnhing IHHD new spets ontinue to e reveled IHID IHPD IHQD IHRF
sn the followingD the in)uenes of vried hlogented dditives on the synthesis of gde
nnorods nd possiilities to exploit themwill dd to this knowledge nd revel new
degree of freedomin nnoprtile shpe ontrolF

PFP gde nnoprtile shpe evolution tuned y hlogented dditives
2.2 CdSe nanoparticle shape evolution tuned by
halogenated additives
o (nd out more out the role of IDPEdihloroethne in the formtion of gde nnopyrE
mids nd the e'et of employing other hlogen ompoundsD syntheti studies with vried
dditives were omined with surfe nd ompositionl nlysis nd omplemented with
density funtionl theory lultionsF gentrl questions to e nswered wereX IA why does
hexgonl pyrmidl morphology form nd PA is there wy to in)uene the shpe deliE
ertelyc efter the exmintions of di'erent spets the shpe evolution will e disussed
sed on the (ndingsF
2.2.1 Effects of 1,2-dichloroethane on the synthesis of CdSe nanorods
sn the high temperture wetEhemil preprtion of gdeEron llotrope ompositesD
residues of IDPEdihloroethne @hgiA were shown to evoke morphologil trnsformtion
of in situ prepred gde nnorods to hexgonl dipyrmidl prtiles @gde pyrmidsA
with wurtzite struture during the ourse of the retion ISD ITF he hlogented solvent
hd entered the retion mixture s dispersnt for the injeted ron llotropes nd ws
in lrge prt removed in vacuoF e similr shpe evolution ould e otined y dding
queous hydrohlori id to the nnorodEgx retionF his method ws modi(ed to
protool without ron llotropesD whih sped up the morphologil evolution IUD IHSF
e hindrne y ron llotropes my e explined y distured di'usion )ux ndGor
dsorptive intertions of neessry omponents with ronF
sn typil gde nnopyrmid synthesisD hgi ws injeted to omplex of dmium
with n-otdeylphosphoni id @yheA in tri-n-otylphosphne oxide @yyA t VH
efore the injetion of selenium in triotylphosphne @yeD I wD injeted t PTS A
initited the formtion of nnoprtiles whih were left to grow t PSS for R hours
@molr rtiosX gdGeGyheGhgi a IXPXPXHFUAF sn pigure PFSD the evolution of the (rst
sorption mximum of smples from synthesis with hgi is plotted together with dt
from ontrol experiment without dditive @the full sorne spetr n e found in
eppendix eAF he fst shift of the sorption during the (rst IH minutes n e orreE
lted with kinetilly ontrolled stge of rod growthF efterwrdsD the shift slew down
nd the diretion of growth hngedF sn the provided trnsmission eletron mirosopy
@iwA mirogrphs the nnoprtiles hve grown perpendiulrly to the cExis with sE
pet rtios pprohing one nd grdully eliminted the )t {IH©IH} side fets to give

Figure 2.5: Temporal dependence of the rst absorption maximum in reactions with and without
1,2-dichloroethane (DCE) and the morphological evolution of nanopyramids with
DCE in TEM micrographs. The aspect ratio of the nanoparticles develops from
around three after 10 minutes to approximately one after 240 minutes.
hemimorphi hexgonl dipyrmids with @HHHIA se nd dominnt {IH©I©I} fets IHTF
prom thisD it n e dedued tht the hnge of growth rtes of the fets is used y
thermodynmilly ontrolled ripening proesses ourring with the depletion of the suE
persturtion t proeeding retion times UWF eent studies on nuletion nd growth
of nnoprtiles suggest tht the seprtion of growth stges is not s ler s desried
erlier due to )ututions in onentrtions IHUF xeverthelessD the urrent retion seems
to e well desried y the fourEstge model with solute formtion @indution periodD sAD
nuletion @ssAD kinetilly ontrolled rod growth @sssA nd thermodynmilly ontrolled
ripening to pyrmidl nnoprtiles @sAF ogether with the sorption fetures the emisE
sion shifted towrds longer wvelengths with timeD ompnied y rodening of oth
due to redution of on(nement with lrger nnoprtile sizes IHS @see spetr in epE
pendix eD ompre to retion with ron nnotues IHTAF e derese of the emission
intensity is noted with proeeding retion timeF pinl quntum yields mesured ginst
hodmine Tq were elow I7 for puri(ed smplesF uh low quntum yields often our
with omprtively lrge gde nnoprtiles nd indite tht the (nl pyrmidl shpe
exhiits high numer of surfe defets @trp sttesA IHVD IHWF

PFP gde nnoprtile shpe evolution tuned y hlogented dditives
sn synthesis without hgiD the sorption wvelength nd with it the size of the
nnoprtiles is shifted to smller vluesF he shpe develops in the wy desried y
eng nd oEworkersD with rods growing oneE nd lter threeEdimensionllyF ipening to
pyrmidl shpes ws not oserved in the ontrol retion even fter UH hours when ystwld
ripening hd rodened the size nd shpe distriution @see eppendix eAF
fy hnging the mount of hgi in the retionD omprison of hgiGgd rtios from
HFQ to IFH reveled two tendeniesF ith inresing mount of the dditiveD @IA the size of
rodsD oth dimeter nd cExisD fter IH minutes eme lrger nd @PA the size of the rods
is relted to the size of the forming nnopyrmidsD in whih the pyrmidl morphology
ws more shrply fetted with more hgiF wirogrphs fter IH nd PRH minutes with
histogrms of the shorter xis @dimeterA fter PRH minutes re depited in pigure PFT
together with the temporl evolution of the (rst sorption mxim during the (rst stge
of growthF he orresponding nnoprtile dimensions re listed in omprison with others
in le PFIF sn Eis spetrD the stge of rod growth ws ompnied y rpid red
shift of the sorption mxim in the eginningF sn ontrst to hgiGgd rtios of HFQ nd
HFUD the solution remined olourless until R minutes fter the injetion of e when rtio
of IFH ws employedF roweverD fst redEshift ourred fterwrds with only minor shifts
fter IH minutesF he slope etween sorption mxim is lso smller with hgiGgd HFU
thn with HFQD whih n e interpreted s fster trnsition to the ripening stge with
higher onentrtions of the hlorinted dditiveF
he dely oserved with the rtio of IFH my e used y n inhiition of prtile
formtion nd fstD less de(ned growth one the nulei were formedF snsted of smooth
{IH©IH} side fets the rods preferentilly exhiited zigzg shped onesD inditing strong
in)uene of hgi on the surfe of growing rods t this hgiGgd rtioF yn ompring the
mirogrphs fter IH nd PRH minutes it eomes pprent tht the irregulr shpeD size
nd size distriution of the initil rods in)uenes the qulity of the lter formed pyrmidsF
ith higher irregulrity of the rodsD the numer of stking fults @visile s stripes
perpendiulr to the cExis of the nnoprtilesAD nd the inhomogeneities in shpe nd
size of pyrmids were more pronounedF ith low mount of hgiD on the ontrryD
tendenies towrds pyrmidl morphology hd developed only in few nnoprtilesD
irrespetive of size nd thus seemingly rndomF he rtio of HFU seems to e su0ient to
promote feting of ll nnoprtiles wheres it is low enough to prevent the formtion of
irregulr rods nd defet loded pyrmidsF

Figure 2.6: (a - c) TEM micrographsafter 240 (large) and 10 minutes(inset) and (d - f) histo-
grams of the diameter of samples after 240 minutes, prepared with dierent DCE/Cd
ratios. (g) Evolution of the rst absorption maxima during early periods of the

2.2 CdSe nanoparticle shape evolution tuned by halogenated additives
Table 2.1: Nanoparticle dimensions at dierent DCE/Cd ratios after 10 and 240 minutes and
with CdCl
instead of CdO as precursor, measured from microgaphs (TEM) or powder
X-ray diractometry (XRD).
DCE/Cd Time
Method c-axis [nm]
Diameter [nm]
10 min
10.4 ± 1.1
4.4 ± 0.4
240 min TEM
12.8 ± 2.0
8.3 ± 1.0
10 min
13.0 ± 1.3
4.7 ± 0.5
240 min TEM
13.1 ± 1.4
12.2 ± 1.3
10 min
16.8 ± 2.6
7.4 ± 0.9
240 min TEM
20.8 ± 2.3
20.9 ± 2.4
240 min XRD
45.8 ± 0.1
52.2 ± 1.6
no nucleation
240 min XRD
34.9 ± 0.5
36.7 ± 0.1
To round o the picture, higher DCE/Cd ratios were tested and the Cd salt (CdO) was
exchanged for CdCl
to be Cd- and Cl-source in one [110]. A ratio of 1.3 evoked a result
similar to CdCl
because in both cases wurtzite CdSe formed but instead of a colloidal solu-
tion bulk like precipitate was obtained. Table 2.1 provides an overview of the nanoparticle
dimensions, with values well above 30 nm for components of the precipitates. A DCE/Cd
ratio of 6.5 was enough to prevent nucleation. These ndings comply with earlier attempts
to use CdCl
or salts of other strong acids as precursors, where the salts were highly soluble
in the reaction mixture but nanoparticle nucleation could not be eected [90, 111]. Accord-
ingly, the solubility of the reaction components seems to be increased by DCEor chloride
and to thus play a signicant role in determining the size of the nanoparticles. Earlier nu-
clear magnetic resonance experiments showed that the Cd-ODPA complex is not observable
under high excess of DCE[16]. Thinking of the LaMer plot, an increase of precursor or
solute solubility shifts the saturation and critical value for nucleation upwards for a given
metal or chalcogenide compound. At the same time, the maximum supersaturation only
moves slightly so that the nucleation window is reduced. A higher concentration of ligands,
for instance, leads to the formation of larger nanoparticles and, due to the inverse propor-
tionality of the number and size of the produced nanoparticles at constant initial precursor
concentration, to a smaller number of nuclei [94, 107, 112, 113, 114, 115, 116]. This line of
argumentation complies with other authors who attributed increased Cd-chalcogenide di-
mensions obtained with higher amounts of halides in solution to a formation of well soluble
Cd-halide compounds or mixed complexes with halides and the original ligands [38, 117].

Figure 2.7: With oleic acid instead of octadecyl phosphonic acid pyramidal shapes evolved under
the inuence of DCE but a high number of tetrapods is visible in the micrograph
after 13 minutes, which results in a mixture of tetragonal and hexagonal morphologies
after 4 hours.
The inuence of DCE on the morphology is not limited to the case with octadecylphos-
phonic acid as ligand, since pyramidal shapes with wurtzite structure were also observed
when oleic acid was employed to complex Cd
. Cd-fatty acid complexes exhibit a higher la-
bility and faster nucleation kinetics than the corresponding phosphonate compounds [118].
In the present reaction, this is reected in the circumstance that the solution turns brown
immediately after the injection of Se, omitting the usual induction time and change from
colourless via yellow and orange colours. In addition to hexagonal structures tetragons are
visible in the micrographs in Figure 2.7, which can be traced backto tetrapodal nanoparti-
cles formed in the rst stage of growth. Tetrapods grow preferably from zinc blende nuclei
[119], indicating that DCE has a general eect on the expression of surfaces diagonal to the
c-axis and the nanoparticle size but ODPA is still the component responsible for anisotropy
and the formation of dened rods.
2.2.2 Shape and size manipulation with other halogen compounds
The circumstance that the amount of DCE inuences the size of nanoparticles similarly to
halides and that the pyramid formation is possible with added hydrochloric acid strongly
suggests that the active species is related to chloride (Cl
) and that dierent additives may
be utilised. It could indeed be shown that the progression of the reaction and thus the
shape evolution can be regulated by the choice of additive. For this, chemical analoga in
form of dierent chloroalkanes and dihaloalkanes with bromine and iodine were compared.
The ligand to Cd ratio had to be increased to 2.7 in this reaction to fully dissolve CdO.

2.2 CdSe nanoparticle shape evolution tuned by halogenated additives Variation of the halogen additives
A variety of chloro components can be applied to induce shape transformations during the
synthesis of CdSe nanorods. Apart from halo alkanes, ammonium salts such as ammonium
chloride and n-dodecyltrimethylammonium chloride (DTAC)are capable of evoking the
formation of nanopyramids and -bullets
, thus supporting the hypothesis of chloride as
active species [110]. There are two major mechanisms thinkable for the release of chlorine
from 1,2-dichloroethane. The rst one is an elimination of Cl
or hydrochloric acid. This
can be thermally activated and is catalysed by transition metal Lewis components so that
process temperatures that are usually above 340 may be lowered to regimes of the present
reaction temperature [121, 122]. Acting as Lewis acids, Cd ions in precursor complexes
in solution and on the surface of already formed nanoparticles might be able to ease the
elimination reaction by "pulling" on Cl atoms in the molecule, while traces of water
as well as other reaction components may act as proton acceptor. Alternatively, Lim et al.
proposed a release of halide through the formation of (R
)-type compounds by
a reaction between the haloalkane and tri-n-octylphosphane. Recently published results of
Cristina Palencia support this latter idea for the nanopyramid reaction [123].
A relation between the molecular structure of the additive and its impact on the
nanoparticle shape could be demonstrated by a variety of haloalkanes. Reactions were
carried out by substituting DCE in equimolar amounts (halo alkane/Cd: 0.7)with 1,1,2-
trichloroethane (TCE), 1,2-dichlorobutane (1,2-DCB), 2,3-dichlorobutane (2,3-DCB) and
1-chlorooctadecane (COD). The additives were injected to the Cd-ODPA complex 10
below their boiling point or in the case of 1-chlorooctadecane (boiling point: 348 )di-
rectly at 265 . Again, aliquots were examined by UV-Vis spectroscopy and TEM as shown
in Figure 2.8 and in Appendix A, where micrographs and absorption data for the full re-
action time are provided. While all reactions yielded pyramidal nanoparticles, dierences
in the temporal evolution of absorption and morphology as well as the homogeneity of
the samples were recorded. The inuence on the nanoparticles and thus the reactivity of
the additives seems to follow a structure related tendency that would be expected for a
substitution reaction. In the beginning of the growth process (after 3 minutes), the wave-
lengths of the absorption maxima are higher and the nanorods bigger (see micrographs in
Appendix A)with additives containing Cl atoms that are less stabilised by neighbouring
alkane groups (+I eect and steric stabilization). Similar to the absorption data shown
The bullet morphology obtained with DTAC might be caused by amine ligands, which can be formed
as side products through its thermal decomposition [120].

Figure 2.8: Temporal evolution of the rst absorption maxima in reactions with chloroalkanes
(1 - 5) and without additive. The employed chloroalkanes are (1) 1,2-dichloroethane
(DCE), (2) 1,1,2-trichloroethane (TCE), (3) 1,2-dichlorobutane (1,2-DCB), (4) 1-
chlorooctadecane (COD) and (5) 2,3-dichlorobutane (2,3-DCB). Longer absorp-
tion wavelengths in the beginning (after 3 minutes) indicate a formation of big-
ger nanoparticles withadditives containing sterically less stabilised chlorine atoms.
Modied withpermission from [124]. Copyright 2014 American Chemical Society.
in Figure 2.5 and Figure 2.6 g, this can be explained by an increased precursor or solute
solubility through released Cl
. The latter causes growth to supersede nucleation at an
earlier point and increases nanoparticle sizes [94]. In terms of the shape evolution dur-
ing ripening, nanopyramids with absorption maxima ranging from 661 nm to 675 nm after
240 minutes were eventually obtained with all additives. With DCE they were most faceted
and homogeneous, whereas more rounded shapes or attened pyramids appeared with other
additives such as 2,3-DCB and COD (micrographs in Appendix A). Changes between the
measured absorption maxima are less pronounced with the large DCE-nanoparticles than
with those prepared with 2,3-DCB, especially in the beginning of the reaction. Between
10 and 240 minutes the shift is 44 nm with DCE and 87 nm with 2,3-DCB. The shifts
with the other additives lay between these values. A smaller shift and the more faceted
morphology of nanoparticles that were large in the beginning indicate a higher tendency to
ripening processes in the monitored period compared to nanoparticles that were smaller in
the beginning. In combination with the results obtained with varied amounts of DCE, it
can be concluded that a higher accessibility or availability of Cl
results in larger nanorods
and more homogeneously sized and faceted nanopyramids.

2.2 CdSe nanoparticle shape evolution tuned by halogenated additives
Figure 2.9: TEM micrographs of samples prepared with DCE (a, d, g), DBE (b, e, h) and DIE
(c, f, i) at dierent stages of the reaction. With heavier halogens a tendency to the
formation of diagonal facets and increased lengths is visible, while the evolution to
pyramidal shapes is less dened or hardly recognisable and the particles exhibit a
high number of stacking faults. Reprinted with permission from [124]. Copyright
2014 American Chemical Society.
The hypothesis of a substitutional release of halide which inuences the extend to which
the shape evolution is aected was further corroborated by experiments with chemical
analogues of DCE containing other halogens, in particular 1,2-dibromoethane (DBE) and
1,2-diiodoethane (DIE). The additives were introduced in equimolar amounts (DXE/Cd
0.7). In the kinetically controlled growth regime, anisotropic nanoparticles terminated with
zigzag shaped side facets similar to the ones obtained with a DCE/Cd ratio of 1.0 evolved.
With both additives the c-axes were even longer (>20 nm) than those observed in the case
of DCE. After 10 minutes (TEM: Figure 2.9 a -c), the size increases from nanoparticles with
DCE to DBE to DIE, which suggests a stronger inuence of the iodoalkane followed by the

bromoalkane during nucleation and early stages of the reaction compared to chloroalkanes.
Thus, the eects on the nanoparticle morphology in this phase of growth must be related
to the reactivity of the haloalkanes, determined by both the molecular structure and the
nucleophilicity or leaving group ability of their halogen atoms.
When approaching the ripening stage, a trend toward pyramid formation is visible
with nanoparticles prepared under addition of DBE, whereas the surfaces of nanoparticles
prepared with DIE are smooth and only a slight tendency of growth perpendicular to the
c-axis is apparent. This means that the heavier halogens boost the growth process during
rod formation but are less eective in the re-shaping process during ripening. The reason
for the latter will be discussed in more detail in the next sections. Size control by additive injection after the nucleation
In order to reduce the size of the nanopyramids, the relation between rod and pyramid size
can be exploited. If the additive is injected after the nucleation and an initial rod growth,
its eect on their dimensions is diminished. Furthermore, the temperature at the moment
of Se injection may be increased, which reduces the critical radius of the nuclei and leads
to even smaller rods. In presence of additives this would be counter-productive because of
an increased release of halide.
An ideal additive for this method is 1-chlorooctadecane with its boiling point well above
the growth temperature of 255 . At an additive/Cd ratio of 0.7, already formed rods with
a c-axis of 5.6 ± 0.7 nm gradually evolved into comparatively small and smoothly shaped
hexagonal pyramids with a c-axis length of 7.8 ± 0.7 nm and a small size distribution,
even after 24 hours. The corresponding micrographs and absorbance spectra are shown in
Figure 2.10. A dierence to the nanopyramids prepared with DCE is that the frustum of
the pyramids is smaller or non-existent (see model in (c)), which reduces the number of
exposed crystallographic vertices and thus reactive sites.
2.2.3 Morphology related changes of phosphorous and halogen
contents examined by surface and elemental analysis
Surface analysis of the nanoparticles was carried out by X-ray Photoelectron Spectroscopy
(XPS) in collaboration with Fabiola Iacono, Roberto Otero and José M. Gallego from
the Universidad Autónoma de Madrid and IMDEA Nanoscience Madrid. XPS spectra
were obtained at the BESSYII synchrotron storage ring, Helmholtz Foundation Berlin-
Adlershof. Employing a synchrotron facility allows for a variation of the photon energy of

2.2 CdSe nanoparticle shape evolution tuned by halogenated additives
Figure 2.10: (a, b) Micrographs of samples before injection of 1-chlorooctadecane and after
24 hours of growth in presence of the additive. The inset in (b) shows enlarged
pyramids with lattice pattern perpendicular to the c-axis. The UV\-Vis absorbance
shifts and broadens onlygradually, before it remains in place after 3 hours.
the monochromatic incident beam, which becomes important when relevant signals overlap
with Auger signals
from other elements. In standard set-ups with Al anodes and a photon
energy of 1486.6 eV, for instance, P 2p signals overlap with LMM Se Auger signals. This
problem is avoided when reducing the photon energy to 620 eV, since the orbital signals
shift but the Auger signals remain unchanged. The examined samples were prepared by in
situ deposition of monolayers onto highly oriented pyrolitic graphite (HOPG) following a
recipe optimised for this purpose [125]. Even though this involved longer reaction times and
slightly changed amounts of ligands, the overall morphological evolution was comparable
to the reactions without carbon substrates.
The three samples examined were nanorods without additive, nanoparticles prepared
with bromine additive DBE and with chlorine additive DCE. Figure 2.11 displays TEM
micrographs of nanoparticles from the supernatant above the HOPG substrates, XPS sur-
vey scans, high resolution P 2p signals of all three samples and high resolution data of
the relevant halogen peaks. It was established that nanopyramids grow in solution and
then attach to the carbon surface [16], so that the shape of the attached nanoparticles is
assumed to be identical with the ones shown.
Auger electrons are generated by ejection from outer shells through energy that is released when an
energetic core electron lls the vacancy left by the electron rst ejected by a photon.

Figure 2.11: (a - c) TEM micrographsof nanoparticle aliqoutsisolated from the supernatant
of XPS samples in situ deposited on HOPG. (d) XPS survey scans of rods and
pyramidsprepared with DBE and DCE (top to bottom) obtained at 620 eV with
an energy pass of 50 eV. (e) High resolution scans of P 2p signals for all three
samples (energy pass 20 eV); the area of the peaks was normalized to the area of
the respective Cd signal. (f) Br 3d orbital signal of the sample prepared with DBE
and (g) Cl 2p signal of the sample obtained with DCE. In (d - g) red lines are the
envelope functionsof the applied tsand the tsare shown asa solid black line.
In (e) the t and envelope functionsare identical.

2.2 CdSe nanoparticle shape evolution tuned by halogenated additives
The Cd and Se XPS signals of the samples did not show signicant dierences, which is
due to the large bulkcontribution from the nanoparticle core, which masks comparatively
small changes on the surface. Judging by the widths and positions of the halogen peaks an
interaction of the atoms with the nanoparticle surface can be assumed. The peakposition
of Cl 2p 3/2 of the sample prepared with DCE is with 198.8 eV close to the value reported
for spherical Cl-capped CdSe nanoparticles and in between the ones in bulkphases of
(198.4 eV) and the more ionic ZnCl
(199.1 eV) [29, 125, 126]. This circumstance
hints at an ionic interaction and supports the presumption that halide ions are the species
adsorbing to the surface and interfering with the shape evolution. Similarly, the Br 3d
5/2 signal, peaking at 68.9 eV, lies in between CdBr
(68.6 eV) and ZnBr
(69.4 eV) [126].
In addition to this, the full width at half maximum (FWHM) of both signals is broader
than the reference signal (Cl 2p 3/2: 1.2 eV; Br 3d 5/2: 1.3 eV; Au 4f: 0.7 eV), indicating
dierent chemical environments for the halogens, for instance caused by the interaction
with dierent nanoparticle facets. A comparison between the P signals, where the areas
were normalized to the area of the corresponding Cd 3d signal, reveals a reduction of the
relative P content in the samples from rods (0.14) to nanopyramids prepared with DBE
(0.09) to nanopyramids prepared with DCE (0.05). Relating this tendency with the above
shown micrographs where the nanoparticles with DCE are of pyramidal morphology while
those with DBE are of a rounder morphology after the same reaction time, leads to the
assumption that the shape evolution induced by halogens is accompanied by a loss of P
Iacono et al. revealed that CdSe nanopyramid attachment to HOPG is promoted in sam-
ples with higher Cl content [125]. There is an inverse correlation of the relative peakareas
between P/Cd and Cl/Cd in rod-shaped and pyramidal samples with dierent Cl content.
In combination with solid state nuclear magnetic resonance spectroscopy and inductively
coupled-mass spectrometry this was attributed to a reduced amount of ODPA-related lig-
ands, especially anhydrides, and incorporation of Cl into the ligand sphere, allowing for a
more intimate contact between nanopyramids and substrates.
Changes of P and Cl contents during the shape evolution in nanoparticle samples pre-
pared after the method without carbon substrate were examined by standardless total re-
ection X-ray uorescence spectroscopy (TXRF) in collaboration with Mauricio D. Coderch
and Ursula A. E. Fittschen from the Institute of Inorganic and Applied Chemistry of the
University of Hamburg. The elemental analysis of nanoparticles and their ligand sphere is
not straightforward. In the present case, the components of interest, P and Cl, are con-

stituents of the ligand sphere and thus only a small fraction of the total amount of atoms
in a sample. There is a limited number of analytical methods capable of determining Cd
and Cl simultaneously under this pretext. For the determination of Cl in the relatively low
concentrations found in the present samples, special systems of inductively coupled mass
spectrometry (ICP-MS) or optical emission spectroscopy (ICP-OES) would be necessary.
Furthermore, the mandatory dissolution of the samples prior to ICP-MS or ICP-OES anal-
ysis leads to problems with a quantitative capture of Cl after acidic digestion. Through the
formation of gaseous hydrochloric acid major amounts of Cl are lost
. In energy dispersive
X-ray spectroscopy (EDX) in combination with transmission electron microscopy (TEM)
light elements are usually detected well but their low content in the samples resided in the
range of the analytical error of the method.
A more convenient and applicable method for elemental analysis of nanoparticles with-
out the need for digestion processes is TXRF. Samples may be deposited onto the sample
holder as a powder or directly from solution. A small drawback is that P and Cl can be
quantied but a satisfactory accuracy in their determination results in larger errors of abso-
lute values for heavier Cd and Se due to re-absorption eects. In addition, the determined
Cd values are too high, which is caused by an overlap of the Cd L-signals used for calcu-
lation by the evaluation software (Spectra) with background Ar. For these reasons trends
but no absolute values should be taken into account in this case. Remarkably, already
aliquots with rod-shaped nanoparticles taken after 10 minutes contain Cl, which conrms
an inuence of the halogen throughout the reaction. A temporal evolution over the course
of a reaction is shown in Figure 2.12. Clear trends in the atomic ratios relative to Cd are
apparent, with an increase of Cl and a decrease of P values with time and thus proceeding
shape evolution. This dependence further supports a mixed coordination of Cd with P
and Cl ligands and is in accordance with the relation found in reference [125]. Despite this
repeatedly observed relation between P, Cl and Cd, the Se to Cd ratio changes without
recognisable trend which must be due to reasons other than the morphological evolution.
An explanation might be the presence of residual precursor material. Even though the
purication procedure was leaned on reported protocols for CdSe nanoparticles obtained
with similar components [78, 127], there might still be varying amounts of impurities in
the analysed samples, calling for a further optimisation of the process for quantitative
elemental analysis.
The same applies for Se, which can partially evaporate as hydrogen selenide. A basic alternative
such as the digestion after Schöninger in an arc lamp and collection of the ashes in sodium hydroxide
solution would be ideal to determine the Cl content. Anyhow, the reaction mixture and nanoparticles
contained several oxygen bound phosphorous compounds which can form explosive mixtures with air
upon intensive heating.

2.2 CdSe nanoparticle shape evolution tuned by halogenated additives
Figure 2.12: Temporal evolution of atomic Cl, P and Se to Cd ratios as measured by TXRF.
The values for Se were divided by 5. Errors arisingfrom sample preparation are
Cl: 9.2%, P: 20% and Se: 6%.
On the other hand, as Anderson et al. formulated, "nanocrystal purity is an inde-
nite concept" [108]. An important question in this context is where the actual limits of
a nanoparticle and its ligand sphere are set. For physical properties, the inorganic core
of the particle and the ligands that ll surface traps will be decisive. For the chemical
reactivity everything that sticks to the nanoparticle during common purication proce-
dures will be important, so that excessive purication will prevent an accurate analysis.
The aforementioned authors proposed that the outer layer of CdSe nanoparticles is com-
posed of ligand coordinated Cd, which improves quantum yields and can be washed away
easily, thus distorting the results. For these reasons and due to incompatibilities of the
samples with many standard work-ups, the analytical description of nanoparticles seems
to be a compromise between necessary purication and determination of the real sample
2.2.4 Ligand-surface interactions and the hexagonal pyramidal shape
On the rst glimpse the pyramidal morphology does not seem to be energetically favourable.
Nevertheless, naturally occurring cadmoselite minerals with hexagonal dipyramidal Habi-
tus were found [128] and the shape is formed at a point in the reaction where a transition
from kinetic to thermodynamic growth control is expected. Together with the fact that

this morphology is polyhedronD whih is likely formed under thermodynmi ontrol @see
setion PFIFPAD it ppers resonle tht hexgonl pyrmids re indeed n equilirium
shpe of the wurtzite struture in the presene of hlidesF o lrify this pointD the lE
redy mentioned experimentl nd theoretil studies of ol' ome to helpF por wurtzite
rystls with ioni or prtilly ioni hrter the ontriution of tioni onds to the
surfe energy is higher thn tht of nionsF husD nion rih fets onstituting hexgE
onl pyrmidl morphologies with lrge @HHH©IA nd {IH©I©I}Etype fets re predited TTF
en inresing degree of ioniity determines whether the pyrmids exhiit frustum with
mixed ions or if they re )t nd ompletely nion rihF uh expeted nionErih pyrmiE
dl rystls were oserved experimentlly for vpour grown wurtzite mterils inluding
gdeD whileD on the ontrryD tion rih surfes with predominnt @HHHIA se formed
in solution y ething with hydrohlori idF he ltter ws sried to n intertion
etween wter moleules or hlogen nd the rystl surfeF his demonstrtes tht the
dsorption of ioni speies exerts signi(nt impt on the surfe energies nd shpe
of rystls nd tht the shpe of gde nnopyrmids is most likely used y gd ound
hlorideF e slight inrese of ioniity nd growth ontrol due to hlide dsorption might
lso explin the ove desried preferene for the wurtzite over the zin lende phse in
other reportsF
o further eluidte the role of di'erent lignds in nnorods nd nnopyrmidsD density
funtionl theory @hpA lultions under periodi oundry onditions were rried
out y ghristin ulinkeF finding 0nities to the dominting fets of hexgonl pyrmids
of possile lignds nd dditives were determined with the yge softwre IPWF wore
spei(llyD the vhe exhnge funtionlD the orreltion funtionl xESQW nd the
ehlrihs sis set were pplied IQHD IQIF he dsorption energies re the di'erE
ene etween the sum of the seprte energies of the gde rystl nd the lignd nd the
totl energy of the omined systemF he gde sls visile in pigure PFIR were kept
t onstnt geometri prmetersD wheres the lignd ws free to relxF he orgni ligE
nds oordinting gde nnorods nd pyrmids re relted to otdeylphosphoni id
IQPD in prtiulr doule deprotonted yhe
nd deprotonted nhydrides with two
or more onneted moleulr units IPSF yther potentil lignds in the retion re triE
nEotylphosphne nd its oxideD s well s the hlolknes with hgi s representtive
nd relesed tomi hlidesF irlier lultions hve shown tht there is little di'erene
etween dsorption energies of short nd long hined lignds WTF por this resonD luE
ltion time ould e sved y employing short lkyl hins @propylphosphoni idD eD

PFP gde nnoprtile shpe evolution tuned y hlogented dditives
insted of otdeylphosphoni idA without hnging the qulittive spets of the reE
sultsF he dt otined with fets most relevnt for gde rods nd pyrmids @depited
in pigure PFIQA is listed in le PFPF
Figure 2.13: hemti depition of rod nd hexgonl pyrmid with distint fetsF
Table 2.2: edsorption energies of di'erent lignd speies on dominnt gde nnorod
nd Epyrmid surfes @eX propylphosphoni idY
@yAX triEnE
edsorption energies e
gde @HHHIA
egd @HHH©IA side @IH©IHA slope gde @IH©I©IA
heomposes upon intertionF
tronger dsorption energies result from the lultions for ll hrged Etype ligE
nds @deprotonted phosphoni id speiesD hlidesAD while neutrl vEtype lignds @yD
yyD protonted phosphoni id speiesD hgiA intert only wekly on ll fetsF
hese (ndings re in greement with previous reports on ligndEgde surfe interE
tions QID UUD UVF ell speies dsor omprtively wekly to the eErih @HHH©IA fetD
wheres they intert more strongly with gdErih sites @see lso UUD WTD IQQAF

por the vEtype ligndsD the highest dsorption energy is found for the nonEpolr @IH©IHA
side fet terminted with oth gd nd e tomsFst is followed y the vlues for the
gdErih @IH©I©IA slope nd @HHHIA ottom fetsFhe lest fvourle intertion ours
with the eErih @HHH©IA fetFhe preferene of these lignds for the side fet on(rms
erlier lultions whih were utilised to explin the growth of nnorods UUF
sn omprison to the Etype ligndsD howeverD the inding 0nity is wekFith the
hrged lignds yhe
nd yheEnhydride
nd hlides the dsorption energy folE
lows di'erent orderFrereD the intertion on the gdErih fets is energetilly most
fvourle nd dereses from the @IH©I©IA slope to the @HHHIA ottomD @IH©IHA side nd
@HHH©IA top fetsFhe lignds most strongly intert with the sloped @IH©I©IA fet euse
this is the roughest one terminted y gd sitesFprom the sl models depited in pigE
ure PFIRD it n e seen tht gd is omptly oordinted y Q e toms on the @HHHIA fet
@ottomAFsn the mixed @IH©IHA side fetD gd is lso oordinted y Q e toms leit in
muh more orrugted fshionFsmportntlyD in the @IH©I©IA slope fetD gd toms hve two
dngling onds insted of oneD sine they re only oordinted y P e tomsFhis leds
to higher tendeny of lignd oordintion nd the high dsorption energies for hrged
lignd speiesD espeilly the dominnt lignd yhe
Fenother remrkle result is tht
protonted @neutrlA phosphoni id nhydrides were unstle on the sloped fet nd
even deomposedFrlides on the ontrry prefer dsorption to this fetFhe energies
derese with the trend of sinking ioniityX gl
b fr
b s
Fking into ount the tenE
deny towrds the pyrmidl shpe oserved in the retionsD it ours resonle tht
the ioniity of the surfe onds is the driving foreD in greement with the ol' modelF
sn reltion with the other ligndsD hlides with their single hrge exhiit dsorption enE
ergies etween the vEtype lignds nd the douleEdeprotonted yhe speiesFith the
oserved dsorption energies nd the ontriutions of ol' it is fesile to lim tht the
pyrmidl shpe is energetilly fvourle under onditions where ligndEsurfe interE
tions nd rerrngement of surfe toms determine the growth rtes of di'erent rystl
fetsFuh onditions re ful(lled under thermodynmilly ontrolled ripeningF
he speil role of gl
might e explined y mixed oordintion of gdErih slope
fets y gl
nd yhe
Fhue to the steri demnd of deprotonted phosphoni id
lignds there is spe enough for the dsorption of smll hlides to dditionl oordintion
sites whih even ulkier nhydrides nnot rehFuh mixed oordintion omplies
with reported trp (lling in smples y hlidesD where the ltter sit in etween or
even prtilly sustitute olei id lignds PUD PVF

2.2 CdSe nanoparticle shape evolution tuned by halogenated additives
Figure 2.14: (a) Side and top-view of PPA
adsorbed to dierent CdSe facets and HR-TEM
images of pyramidalCdSe nanoparticles, (b) top and (c) side-view with indicated
facets. Atoms in beige: Cd, orange: Se. Reprinted with permission from [124].
Copyright 2014 American ChemicalSociety.

2.2.5 A closer look at the shape evolution process
he presented results led to the ssumption tht hlogented dditives 'et the morphoE
logil evolution in onseutive kineti nd thermodynmi growth stges in di'erent wysF
his n e understood y ringing together experimentl nd theoretil disoveriesF sn
pigure PFIS vwerEtype plot of the retion with the two most importnt in)uenes of
the dditivesD the inrese of rod size nd the tendeny to reshpingD is shownF
Figure 2.15: Morphological evolution of CdSe nanoparticles to pyramids under inuence of halo-
genated additives in connection with dierent modes of growth control. Adapted
with permission from [124]. Copyright 2014 American Chemical Society.
Rod growth
foth in retions with nd without dditivesD the growth rte in the diretion of the cExis
is the highest during the (rst stge of growthD whih n e relted with the low dsorption
energies of lignds to the @HHH©IA fet @see le PFIRAF he other fets re tightly pped
y mixture of yhe
nd deprotonted nhydrides whih n ridge djent gd toms
nd isolte the surfe from the solutionF Gpr he e'et of hlorine ontining dditives
with rtios to gd of up to HFU in this stge of growth is n inrese of preursor ndGor solute
soluility nd thus of the rod sizeF he extent of this is more visile with lrger mounts
nd dditives with sterilly more essile ron entres ound to hlorine tomsF his
tendenyD together with the possiility of induing the shpe trnsformtion with hloride
ompounds nd the signs for ioni intertions etween gl nd the nnoprtile surfeD
strongly suggests tht relesed gl
Etype lignds re the responsile ingredientF
et hgiGgd rtio of IFH nd with IDPEdiromoethne nd IDPEdiiodoethne t hiGgdX
HFU the dditive does not only inrese the nnoprtile size ut lso 'ets the roughness of
the side fetsF igzg feting is oserved in prllel to the cExis @setion PFPFPFIAF imilr
nnorod shpes were otined in experiments with lignd mixtures of hexylphosphoni id


Type of Edition
ISBN (eBook)
ISBN (Softcover)
File size
39.8 MB
Publication date
2014 (June)
metal-semiconductor halogen Colloidal chemistry Shape inducing additive Chemie Nanopartikel Nanoparticle

Title: Metal-semiconductor hybrid nanoparticles: Halogen induced shape control, hybrid synthesis and electrical transport
book preview page numper 1
book preview page numper 2
book preview page numper 3
book preview page numper 4
book preview page numper 5
book preview page numper 6
book preview page numper 7
book preview page numper 8
book preview page numper 9
book preview page numper 10
book preview page numper 11
book preview page numper 12
book preview page numper 13
book preview page numper 14
book preview page numper 15
book preview page numper 16
book preview page numper 17
book preview page numper 18
book preview page numper 19
book preview page numper 20
book preview page numper 21
book preview page numper 22
book preview page numper 23
book preview page numper 24
book preview page numper 25
book preview page numper 26
book preview page numper 27
book preview page numper 28
book preview page numper 29
book preview page numper 30
book preview page numper 31
book preview page numper 32
book preview page numper 33
book preview page numper 34
book preview page numper 35
book preview page numper 36
book preview page numper 37
book preview page numper 38
book preview page numper 39
book preview page numper 40
212 pages