Efficient Light Emitting Devices Utilizing CdSe(ZnS) Quantum Dots
in Organic Host Matrices
by
Seth Alexander Coe
Sc.B., Electrical Engineering (1999)
Brown University
Submitted to the Department of Electrical Engineering and Computer Science
in partial fulfillment of the requirements for the degree of
Masters of Science in Electrical Engineering
BARKER
at the
MASSACHUSETrS INSTITUTE
Massachusetts Institute of Technology
OF TECHNOLOGY
May 2002
LIBRARIES
@ Massachusetts Institute of Technology 2002. All rights reserv ed.
Author....
.......................
............................
Department of Electrical Engineering and Computer Science
May 10, 2001
A
C ertified by ..........
.......
.....................................
Vladimir Bulovic
Professor of Ele Arical Engineering
Thesis Supervisor
Accepted by..............
Chairman, Department Committee on
Arthur C. Smith
TQt-ATmJT
Lrsjc
-s-
OF TECHNOLOGY
-1 -
MAY 2 5 2004
LIBRARIES
Table of Contents
I.
Abstract
II.
Background
a. OLED Technology
i. Physics
b. QD Technology
i. Physics
III.
Experimental
IV.
Results
V.
Discussion
VI.
Growth System
a. Transfer System Design
b. MiST Design
VII. Future Work
VIII. Conclusions
-2-
Efficient Light Emitting Devices Utilizing CdSe(ZnS) Quantum Dots
in Organic Host Matrices
by
Seth Alexander Coe
Submitted to the Department of Electrical Engineering and Computer Science
On May 15, 2002
in partial fulfillment of the requirements for the degree of
Masters of Science in Electrical Engineering
Abstract
We demonstrate efficient electroluminescence from thin film structures
containing core-shell CdSe(ZnS) quantum dots dispersed in molecular organic host
materials. In the most efficient devices, excitons are created on the quantum dot sites via
energy transfer from organic host molecules, and direct charge injection into the quantum
dots is minimized. For quantum dots with core diameter 38 A, the electroluminescence
spectra peak at 562nm and have full width at half maximum as narrow as 32nm.
Saturated color devices have external quantum efficiencies as high as 0.61% at the
current density of 7mA/cm 2 . At 125mA/cm 2, the device luminance is 1900cd/m 2, which
corresponds to a luminescence efficiency of 1.5 cd/A. The yield over hundreds of
devices is greater than 90%, indicating a robust material system.
Thesis Supervisor: Vladimir Bulovic
Title: Professor of Electrical Engineering
-3-
II.
Background
Much effort has been devoted to enhancing the performance of organic light
emitting devices (OLEDs) [1] by increasing their efficiency [2,3,4], narrowing [5] or
broadening [6] their emission spectra, or polarizing their emission [7]. In this work we
examine the benefits of incorporating other material systems within organic host matrices
to generate efficient hybrid organic/inorganic LEDs of saturated color. Specifically, we
describe OLEDs incorporating CdSe nanocrystallites, or quantum dots (QDs), as
emissive centers. Integration of inorganic QDs into thin film LEDs has previously been
demonstrated [8,9], however these pioneering studies generated devices with low power
efficiencies.
In this letter we demonstrate that through control of the exciton
recombination region within multilayer hybrid structures, one can improve LED
performance by a factor of 25, reaching efficiencies comparable with state of the art
OLEDs.
a. OLED Technology
Since the late 1980's much work has focused on the use of organic materials for
optoelectronic applications. The pioneering works were by C.W. Tang at Kodak [10,11],
demonstrating the first efficient organic light emitting device, and organic solar cell.
Since this time, many academic and industrial groups worldwide have added to both the
list of applications of organic materials, and to our understanding of the underlying
physical processes that are prevalent in organic semiconductor systems.
Optically
pumped lasers [12] and 75% quantum efficient photodetectors [13] have since been
-4-
created, and OLED devices can now be realized in their transparent [14], flexible [15],
metal-free [16], or phosphorescent [17] forms.
Concurrently with all of these research advances, organic materials have also been
introduced into the commercial sector. OLED displays are available on commercial cell
phones and car stereos. Full color flat panel OLED displays with a 13" diagonal have
been demonstrated, though they are not yet available for commercial consumption. It is
clear that commercial organic optoelectronics will be widespread in just a few years time.
However, it is important that university research continue to push the envelope of what is
possible using organic materials. As our understanding of their basic physics grows, new
device structures are conceived, and efficiencies of the old device structures constantly
improve.
It has become clear in recent years of OLED development that there are several
basic physical problems that will be the most difficult to overcome. One is the relatively
short lifetime of any EL device. Lifetimes have been extended using new chemistry to
prevent crystallization, and more sophisticated packaging schemes to prevent water and
oxygen from degrading the emitting molecules. However, the packaging methods are
difficult to extend to flexible substrates, and the chemistry optimization must be repeated
with each new emitter (minimum of three for a full color display) that goes into the
device.
Such constraints have prevented the development of a high efficiency blue and
red OLEDs that have lifetimes long enough to make commercial displays.
Human
perception of luminous intensity peaks sharply in the green, making blue and red devices
much more demanding to create at the same efficiency. In addition, broad emitting blue
-5-
devices will either look green-blue to the human eye, or emit a large amount of UV light,
which we cannot perceive at all.
Thus, creation of efficient emitters at the spectral
extremes requires a narrowband source. Organic molecules tend to have steric flexibility
(as compared to a covalently bonded inorganic crystalline semiconductor), which
functions to broaden the characteristic emission of an amorphous solid relative to a
crystal. While attempts have been made to use chemistry to "bridge" some of these steric
freedoms, and hence create more rigid, narrowband emitters, the attempts have met with
only limited success [18]. The problem is basic to the material set, and hence we must
look outside of the world of organic materials to find solutions.
i. Physics
Fig.1 shows a typical heterostructure OLED. The indium tin oxide (ITO) anode
injects holes into the hole transport layer (HTL), which in this case in the small molecule
N, N'-diphenyl-N, N'-bis (3-methylphenyl)-(1 ,1'-biphenyl)-4,4'-diamine (TPD). Holes
are transported in the TPD until they reach the tris-(8-hydroxyquinoline) aluminum
(Alq 3) interface, where they buildup at the band edge mismatch.
Simultaneously,
electrons are being injected into the Alq 3 from the cathode which is a low work function
metal, in this case Mg:Ag. These electrons are transported through the Alq 3 to the same
heterojunction, where they too accumulate. Some leakage of either holes or electrons can
occur at this interface, allowing the creation of tightly bound electron-hole pairs on
individual molecules of either TPD or Alq 3. These e-h pairs, or excitons as they are
called in the organic literature, are then free to diffuse or energy transfer as single
particles.
They can also relax either radiatively, emitting light characteristic of the
bandgap of whichever material that they were residing on, or nonradiatively, losing the
-6-
energy to phonons or heat. In this TPD/Alq 3 device, all of the excitons will transfer their
energy to the Alq 3 molecules prior to relaxing, and hence no TPD emission will be
observed. Such a device may be made as high as 1% externally quantum efficient, which
is to say that for every 100 electrons injected into the device, one photon may be
observed by a detector external to the device.
To improve the LED luminescence efficiency, the exciton-generating region is
often doped with dyes of high PL quantum efficiencies [19]. The host excitons can then
transfer their energy to luminescent dopants. Similarly, the energy level structure of the
dopants might render them an electron or a hole trap for mobile carriers, which can also
Even small doping concentrations (from a
lead to exciton formation on the dopant.
fraction of a percent to a few percent) in the electrically pumped host layer are sufficient
to change the luminescence spectra, often completely quenching the host luminescence.
Low dopant concentration implies a large average distance between these molecules,
which reduces exciton quenching from dopant-dopant interactions and yields highly
efficient luminescence. The lifetime and efficiency of doped OLEDs have been found to
depend on the nature and concentration of dopant molecules in the light emitting layer
[20].
In Fig. 2 we show some examples of doped Alq 3 OLEDs. For comparison, the
luminescence spectra of undoped Alq 3 is also shown.
Frster energy transfer in
Alq 3:(2%) DCM2 ([2-methyl-6-[2-(2,3,6,7-tetrahydro-1H, 5H-benzo [ij] quinolizin-9-yl)
ethenyl]-4H-pyran-4-ylidene]propane-dinitrile) OLEDs leads to orange EL with forward
external emission efficiency of qexffoard~) 0.5% photons per injected carrier [19,21].
Dexter energy transfer in Alq 3 :(6%) PtOEP (2,3,7,8,12,13,17,18-octaethyl-21H,23Hporphine platinum (II)) OLEDs leads to highly efficient, 77x(forward)
-7-
4%, saturated red
phosphorescence [22]. Also, coumarin 6 doped in Alq 3 improves the color saturation and
efficiency (rex(foward)
2%) of green OLEDs [19], due to carrier trapping. In these doped
OLEDs, different molecules perform the functions of electron/hole transport, and of the
luminescent center. Such separation of functionality allows for optimization of device
performance.
Inorganic nanodots, such as those of CdSe (see Fig. 3) are ideal dopants in
organic host materials. Their photoluminescence efficiency is comparable to the best of
organic lumophores, while the spectral emission width is less than half of that of a typical
organic emitters, resulting in more saturated colors. CdSe nanodot emission can be tuned
across most of the visible spectrum by changing the dot diameter from 17 to 120A (see
Fig. 3). Such wide tunability by simple size selection greatly simplifies optimization of
QD-LED color emission (an advantage not easily achievable with organic lumophores).
b. QD Technology
Luminescence properties of inorganic QDs match the performance of good
organic lumophores. For example, nanocrystallites of CdSe coated with a ZnS
passivation layer have solution photoluminescence (PL) quantum efficiencies of as high
as 50% [23].
In addition, by changing the diameter of the CdSe core from 23 to 55A, the
peak luminescence wavelength can be precisely tuned from X=470 to X=640 nm with a
typical spectral full width at half of maximum (FWHM) of less than 40nm [24]. This
narrow spectrum is perceived as saturated color, motivating the development of efficient,
saturated color, quantum dot light emitting devices (QD-LEDs). We expect that even red
and blue efficient QD-LEDs can be generated, since in narrow-spectrum LEDs no
photons are lost to infrared and UV emission.
-8-
The broadly tunable, saturated color
emission of quantum dots is unmatched by any class of organic chromophores.
Furthermore, environmental stability of covalently bonded inorganic nanocrystals
suggests that device lifetimes of hybrid organic/inorganic LEDs should match or exceed
that of all-organic LEDs, when QDs are used as luminescent centers.
A quantum dot can most generally be defined as a material of nanoscale in all
three dimensions such that its electronic structure is modified quantum mechanically.
They are also frequently referred to as nanodots and nanocrystals. Recent advances in
the synthesis of such materials by the Bawendi group at MIT have enabled the
incorporation of this unique material set into useful devices.
The synthetic route of
Murray et al. departed from the more standard vacuum MBE growth of randomly
arranged QDs on a perfect crystalline substrate. Instead, this new technique allowed for
the epitaxial growth of single crystals out of solution. This enhanced the level of control
that researchers had in preparing the material, and simultaneously allowed the QDs to be
handled as if they were organic "macromolecules" in solution, since they were grown
with an organic capping layer already assembled on their surface. QDs prepared in this
manner can have extremely narrow emission and absorption bands due to the quantum
confinement effects and the monodispersity of a good sample. In addition, they can be
extremely efficient emitters, with photoluminescence quantum efficiencies as high as
50%. And finally, their emission can be tuned continuously through large portions of the
spectrum. The CdSe quantum dots that we concentrate on in this work emit anywhere
from 470nm to 630nm light, almost the entire visible spectrum.
In contrast to OLEDs, QD technology has not yet reached the commercial
markets.
There is much interest however, in utilizing QDs in such applications as
-9-
combinatorial chemistry and narrowband emitting sources for communications.
While
these applications may not be realized for many years, they are likely to be the first
instance of low-dimensional physics impacting the lives of consumers.
From our perspective, QDs can be seen as the ideal emitting body for
incorporation into OLEDs. Their emission can be tuned throughout the visible spectrum,
and thus they can become the only emitting body that is necessary in a full color flat
panel display.
In addition, their emission is extremely narrowband, allowing for the
creation of highly luminous efficiency LEDs in the blue and red, where otherwise light
would be lost to the ultraviolet and infrared respectively. As an inorganic emitting body,
they have the potential to be far more stable than any organic molecule, as their
degradation mechanisms are far more understood. Additionally, they still represent an
ongoing and fascinating topic of research, and it is our hope that in incorporating the QDs
into LEDs, we will be able to contribute to the body of knowledge that is currently being
accumulated on how the QDs operate on the most basic levels.
i. Physics
Understanding of how a QD functions requires at least a modest understanding of
the principles of quantum mechanics and solid state semiconductor physics.
In a
conventional bulk semiconductor, the charged species (electrons and holes) may be best
described by particle wave functions in a periodic medium.
However, these
wavefunctions are unconfined and experience the periodic medium throughout their
extent (which is typically on the order of 1 Onm). In a confined system such as a QD,
these wavefunctions are not able to extend fully, and hence exist at some raised energy
state relative to the bulk case. Quantum mechanics allows us to consider the relative
-10-
energies of a particle within such a confined system, but intuition is sufficient to tell us
that this energy is raised as the level of confinement is raised (Fig.4). In the case of a QD,
the confinement is determined by the particle size, and hence as the QD diameter is
reduced, the energy that its charged particles have is similarly increased. Thus, a relaxing
electron-hole pair will release more energy (bluer light) as the QD diameter is decreased.
The upper size limit at which this effect exists is the diameter at which a bulk exciton
would occupy, which for the case of CdSe is about 12nm. At extremes of smaller size,
surface effects dominate over any effect of the core crystal, and hence establish some
minimum diameter (near 1.7nm) of a functional CdSe QD.
The function of the ZnS overcoating is similarly explained by considering surface
states. All of the above mentioned wavefunctions rely upon the structure being perfectly
periodic.
A perfect single crystal of CdSe has this periodicity throughout its bulk.
However, this necessarily breaks down at any surface do to the crystal structure itself.
Almost all known inorganic semiconductors are covalently bonded solids, with four
bonds per atom. At the surface of a crystal, the outermost atoms do not have neighbors to
bond to, generating so called dangling bonds.
Even perfect crystals undergo some
surface rearrangements to minimize the energy of these dangling bonds. The crystal
periodicity is thus broken near the surface, giving rise to surface states, which are usually
within in the bandgap of a semiconductor. In the case of a CdSe QD, these surface states
lead to non-radiative relaxation pathways, and thus inefficient emission. The effect is of
course made more pronounced the smaller the QD diameter, since the surface to volume
ratio approaches infinity.
- 11 -
ZnS overcoating mitigates all of these effects.
By adding another shell of
semiconductor to the crystal, the CdSe dangling bonds are passivated. They are instead
replaced with ZnS dangling bonds at the new surface. ZnS also had a larger bandgap
than the CdSe, and hence the electronic wavefunction is still well confined within the
core. Thus, the wavefunction has only minimal overlap with the dangling bonds, and the
surface states, and gives rise to a higher quantum efficiency QD.
The extremely narrow emission band that we are relying upon for our narrow
emission QD-LEDs is also due to quantum confinement effects.
In a bulk three-
dimensional semiconductor, the density of available states above the conduction band rise
very quickly (g(E)ocE 112 ). However, as the dimensionality of the crystal is reduced, the
functional form of the density of states changes. For a truly zero-dimensional crystal the
density of states becomes a series of delta functions, one at each of the confined levels
with in the "box".
In the physical system of a QD between Inm and 12nm, these delta
functions become slightly spread out, giving rise to the bandwidth of QD emission that
we are accustomed to observing. Thermal effects contribute to this broadening as well,
as does the size dispersity in any individual sample of QDs.
III. Experimental
In this study, we incorporate inorganic quantum dots into electrically pumped
molecular organic structures to prove their efficacy in LEDs and examine the
mechanisms of their electroluminescence (EL).
Our basic device structures are shown
inset in Fig.5, along with a schematic drawing of a core-shell type QD passivated with
trioctylphosphine oxide (TOPO) caps. Solutions of QDs in chloroform were prepared by
the synthetic technique of Murray et al [25]. Solution luminescence and absorption
12-
spectra peak at X=561nm and X=548nm respectively, corresponding to a CdSe core
diameter of approximately 38A overcoated with 1.5 monolayers of ZnS.
The QD
solution PL efficiency is 20-25%. The QDs are mixed in various concentrations into a
chloroform solution of TPD, which is then spin-cast onto clean, ITO-coated glass
substrates, resulting in a 40nm thick film. For device I, a 40nm thick film of Alq 3 is then
thermally evaporated onto the TPD:QD layer, and capped by a 1mm diameter, 75nm
thick Mg:Ag (10:1 by mass) cathode with a 50nm Ag cap. For devices II and III, a 5nm
layer of TPD and a 10nm layer of 3-(4-Biphenylyl)-4-phenyl-5-tert-butylphenyl-1,2,4triazole (TAZ) is evaporated prior to Alq 3 deposition (see device structures in Fig. 1). The
spin-casting and device manipulation during growth is performed in a dry nitrogen
environment, with moisture and oxygen content of less than 5ppm. All measurements are
done in air. Device yields over hundreds of devices are greater than 90%, indicating a
robust material system.
The choice of organic host for the QDs is limited by material deposition methods.
CdSe QDs are typically arranged into thin films by spin-casting from solution.
While
spin-casting is possible for molecular organics, and typical for polymer organics, it limits
the available organic matrix materials to those that are highly soluble in solvents such as
toluene, hexanes and chloroform, which are the preferred solvents for the TOPO capped
QD colloids.
In order to have a large range of possible solution mixtures and film
thicknesses, it is necessary to have organic solubility in the range of 10mg/mL. Such is
the case for TPD in chloroform. TPD has the added advantage of being a blue emitting
material, which can facilitate access to the entire visible spectrum by doping different
sized QDs into this organic matrix. Absorption measurements indicate that QDs make up
- 13
-
only 5% by volume of our 400A films (see Fig.5). This corresponds to a layer that is
20A thick.
Considering that QD cores are 38A in diameter (-50A including the
overcoating and organic caps) they can at most form a partial monolayer, and are more
likely dispersed within a complete layer. This is true even if the QDs completely phase
segregate from the TPD during the spinning process. The large ionization potential of
QDs (Fig.6) prevents them from trapping holes in TPD, which is the motivation for
incorporating QDs into the hole transporting layer of these devices. The spectrum and
efficiency of QD-LEDs strongly depends on QD concentration in the TPD matrix. For
low concentrations of QDs the QD-LED behavior is similar to an undoped device, and at
extremely high QD concentrations we observe a morphology change in the QD doped
layer that leads to poor device performance and low yields.
IV.
Results
Fig.7 shows the device structures and electroluminescence spectra of three QD-
LEDs. The inset cartoon depicts the QD core-shell-cap structure.
Spectral peaks at
X=562nm, at X=400nm, and the broader shoulder centered at X=530nm we attribute to
QDs, TPD, and Alq 3 emission respectively. The dashed lines show the decomposition of
the EL spectra into Alq 3 , TPD and QD contributions, which we describe in terms of an
emission fraction (fx).
In device I, holes are injected from the ITO contact into the TPD host matrix and
are transported towards the junction. Similarly, electrons are injected from the Mg:Ag
cathode into the Alq 3 and are transported to the heterojunction. Electrons are trapped at
the QDs due to the relative energy alignment of the lowest unoccupied molecular orbital
(LUMO) levels of TPD, Alq 3 and the QDs. For these charged QDs the barrier to hole
-
14
-
injection from the TPD is greatly reduced. Upon acceptance of holes from TPD, excitons
form on the QDs, and can subsequently recombine radiatively. Alternatively, excitons
can be formed on organic sites near the TPD/Alq 3 interface, which can then energy
transfer to the lower energy QD sites, where they recombine radiatively.
The external quantum efficiency of device I exceeds i1=0.5% for a broad range of
device luminances (from 5 to 1900 cd/m 2 ), peaking at il=0.61% at 21mA/cm 2 (see Fig.8).
In analyzing the external quantum efficiencies for these devices, it is important to
differentiate the organic contributions from the QD component. A linear superposition of
the two components describes the system:
(Eq. 1)
11Totaf flQD"fQDe7lAlq3 fAlq3
We use
IlAIq3=0.
9
%
as determined from experiment, and find that I1QD=0. 4 3 % for device
I. The brightness of 100cd/m 2 is achieved at I=5mA/cm 2, V=6.lV.
characteristics for all three devices is shown in Fig.9.
Current voltage
At 125mA/cm 2, the LED
luminance is 1 900cd/m 2 , which corresponds to a luminescence efficiency of 1.5cd/A, and
a 25 fold improvement over the best previously reported QD-LED result [26].
The thickness of the TPD:QD layer plays a critical role in determining the device
properties.
With a 70nm thick TPD:QD layer, the Alq 3 emission is completely
suppressed at the expense of lower quantum efficiency and higher turn-on voltage of the
device. Thinning the TPD:QD layer, however, leads to an excess of hole injection, and
thus enhanced Alq 3 emission.
Device H (Fig.7) that integrates an additional 50A of thermally evaporated TPD
under the Alq 3 . This TPD layer ensures that almost no excitons are created by direct
-
15
-
charge injection in QDs, since the electron transport in TPD is limited. The observed QD
emission (fQD=0. 3 6 ) is therefore primarily generated after exciton energy transfer from
TPD to QDs. As a consequence i1QD=0. 76 % (calculated from Eq.1), almost a factor of
two better than in device I. Indeed, in device I the emitting QDs were collocated with a
buildup of electrons at the TPD/Alq 3 interface, facilitating the creation of negatively
charged QDs. Charged excited QDs are non-emissive [27], as they may relax through
fast (t= l00ps), non-radiative Auger recombination. Therefore, by moving the QDs away
from the electron buildup at the interface of our devices, we reduce QD charging and
increase QD EL efficiency.
Device III includes 1 OOA of a hole and exciton blocking material, TAZ, under the
Alq 3 layer. Here the Auger process once again reduces the QD EL efficiency, and hence
T1QD=0.
3
1%. The narrow emission spectrum is almost completely due to QDs
(FWHM=32nm).
There is almost no emission from the Alq 3 (fAlq3=00
9
)
due to the
effective blocking of both holes and excitons from transport into the Alq 3. There is some
TPD emission (fTPD=0.0 6 ) from this device due to incomplete energy transfer from the
TPD molecules to the QDs. In addition, there is minimal QD deep trap emission
DT
0-0 1), which
(fQD-
would be observed in the near infrared region of the spectrum [23]. This
deep trap emission is enhanced when incorporating core only QDs, but 1-2 monolayers of
ZnS overcoating is sufficient to eliminate this infrared emission tail (see discussion
section for more detail). A summary of all of these excitonic processes for the three
devices is shown in Fig.10.
-16-
V.
Discussion
The fundamental limits of QD-LED performance are different than those of
OLEDs. For example, in this study QD-LEDs have an emission FWHM as small as
32nm. In contrast, typical molecular organic LEDs have a FWHM of between 60 and
100nm, although emission of some polymers and phosphorescent molecules was shown
to be as narrow as 26nm FWHM [28,29]. The vibrational structure of sterically flexible
organics typically generates broad single molecule emission spectra at room temperature
[30]. The same is not true of the rigid, covalently bonded inorganic QD, for which single
QD spectroscopy shows that the fundamental FWHM linewidth of a QD at room
temperature is 14nm [31]. It is the combination of spectral diffusion and size distribution
of QDs in a sample that yields further line broadening.
Consequently, our 32nm
linewidth corresponds to a size distribution of about 8%. It is reasonable to expect that
current techniques in QD preparation and processing could yield QD-LED line widths
that are as narrow as 25nm.
This true color saturation would be ideal for many
applications where efficient production of narrowband light is desired. In particular, the
creation of high luminous efficiency red and blue LEDs requires both high external
quantum efficiency as well as narrowband emission, to prevent the bulk of emission from
occurring in the infrared or ultraviolet, respectively, where our eyes have minimal
response.
The semi-log plot of the spectral emission intensity as a function of drive current
is shown in Fig. 11 for (a) ZnS coated CdSe nanodots and (b) uncoated CdSe nanodots.
Chemical reaction for the coating reaction is indicated in the inset. In both cases dots are
surrounded by organic ligands to facilitate solvation in the organic solvent during the
- 17
-
spin-on process.
Both plots in Fig. 11 show a strong luminescence peak at 570 nm
corresponding to the nanodots emission, but we also observe a long luminescence tail
from 700 to 950 nm. This emission is due to the luminescence from deep traps (mid-gap
states) caused by dangling bonds on the surface of the nanodots. Trap luminescence is
very inefficient and is more pronounced in uncoated dots.
This typically results in
several-fold lower efficiency of uncoated devices as has been previously observed in the
PL spectra of the nanodots [32]. The progression of luminescence as a function of device
current indicates that for the coated-nanodot devices, Fig. (a), luminescence of deep traps
stays constant as the luminescence form the nanodots increases with current.
For
uncoated-dot devices, deep trap luminescence rises with current together with the dot
luminescence. This is as expected: the small number of defect levels in the coated-dot
devices is first populated at lower currents and then the luminescence form the confined
states of the dot can start as the current is increased. The number of defect states on
uncoated dot devices is significantly larger and cannot be filled completely even at higher
device currents, so its luminescence steadily increases with the increase in current.
Fig. 12 shows the linear-linear plot of the data from Fig. 11(a) where in Fig. 12(b)
spectra are normalized to the intensity of 570 nm peak. From Fig. 12(b) we can see that
500 nm shoulder increases in intensity as the current increases. This peak is due to the
Alq 3 layer.
The imbalance of charge injection and the difference in luminescence
lifetimes of nanodots and Alq 3 molecules then gives rise to the increase in the 500nm
shoulder as a function of the injected current.
We do not expect ZnS coated CdSe nanodots to significantly participate in the
conduction processes of QD-LEDs as the insulating ZnS cap would limit direct carrier
- 18
-
injection into the nanodots. However some amount of tunneling is expected due to the
thin cap. We are investigating if we can use F6rster energy transfer of exciton energy
form the organic host molecules to the nanodots as this would not be affected by the cap
presence as long as we are within the Forster energy transfer radius. We expect the
Fdrster energy transfer radius to be on the order of 40A due to the high absorption
coefficient of nanodots (> 105 cm 1 ). Furthermore, ZnS cap should prevent nanodot
quenching due to aggregation, further alleviating concentration quenching problems
encountered with typical organic dopants.
VI.
Growth System
Equipment needs have been central to the successful completion of this project.
Starting the work in a new laboratory this was necessarily the case, but was further
complicated by the constraints inherent in making QD-LEDs.
Furthermore, at the
projects onset, we could only guess as to what those constraints might be, and as such
may have erred on the side of over designing the system. From prior experience in
OLED development, we are very aware of the importance of isolating water vapor and
oxygen from our devices. We assumed that this would also be the case for the QD-LEDs,
since QDs in neat films are known to degrade quickly if exposed to air. As such, we set
out to develop a complete growth system (see Fig. 13) that would allow the creation of a
multiplayer organic and QD device without exposure to air. In addition, it was desirable
to have as little exposure to atmospheric pressure as possible.
The necessary components of such a system are a spin-coater for solution
processing, and area for device manipulation, some allowance for shadow masking, and a
thermal evaporator. In addition, we knew that an ideal system would include a new
-19-
method of vacuum depositing QDs (see Fig. 14), a sputterer for production of transparent
cathodes, and a probing station for vacuum testing of our completed devices. Integrating
such a large set of chambers into one system was a clear challenge, and led us to spend a
significant time designing the backbone of the system, the transfer line. In addition,
considerable effort has been expended on the creation of a Misted Solution Transport
(MiST) deposition chamber that we hope will solve the problem of vacuum deposition of
QDs, as well as a wide variety of other materials that heretofore have only been solution
processed.
The completed growth system will integrate the method for physical and vapor
phase deposition of hybrid organic/inorganic thin-films with a low-pressure RF/DC
sputtering chamber, an evaporative growth chamber, and a chemical vapor deposition
chamber.
The completed integrated growth system will be capable of depositing
molecular organics, polymers, metals, metal oxides, inorganic nanodots, and colloids in a
controlled layer-by-layer fashion.
An in-situ shadow masking system will enable
fabrication of complex patterned structures inside a vacuum environment, while the
integrated N2 -filled, dry glove box will facilitate handling, measuring, and packaging of
organic thin film samples that are susceptible to reactions with atmospheric oxygen and
water vapor.
Completed samples will be in-situ tested in the analysis chamber by
contacting them with an electrical probe attached to an X-Y-Z manipulator. Optical ports
on the chamber allow for a telescopic view of the devices and facilitate optical excitation
of probed samples. Optoelectronic properties of the hybrid materials and structures will
be investigated at a range of temperatures form 5 K to 600 K, generated by the liquid
helium cryostat and the boron-nitride heater situated behind the sample stage.
-20
-
The
AFM/STM chamber will facilitate in-situ atomic scale microscopy for evaluating
morphology and electronic properties of hybrid materials.
All chambers are connected to the central transfer system that has linear degrees
of freedom. Each chamber is isolated form the others by a gate valve resulting in typical
base and operating pressures as indicated below. Maximum substrate size is 10 cm with
a 5% variation in the thickness and composition of deposited films over the substrate
area. Presently the system consists of the double glove box, evaporator, the sputtering
system, and the transfer line with the load lock.
Operating
Base Pressure
Pressure
SYSTEM
[Torr]
[Torr]
PVPD
10-8
1 to 10
Sputtering
10-9
10-1 to 10-2
Evaporative
10~10
10~10
CVD
10-8
1 to 10
10-9
1 0 -7
Probe Station
1010
10-10
AFM/STM
10-10
10~10
Load Lock
10-7
Central
to 10-
Transfer
10-7
atmosphere
-
21
-
to
a. Transfer System Design
With our need for a sophisticated transfer system well established, we started to
establish its design criterion. The line would need to fit within the space available to us,
integrate at least five chambers as well as a loadlock to a glove box, and provide a means
for in-situ shadow masking. Finally, the system should be designed for ease of use, and
as fool-proof as possible, to maximize up time and minimize the skill level needed to be a
user of the facility.
The method of handling the typical fragile shadow masks within stainless steel
vacuum chambers was a challenge in and of itself. We will summarize our solution to this
challenge with a picture of the end result, Fig.15.
This two piece stack allows the
separate handling and storage of shadow masks, which can then be self-aligned within
ultra-high vacuum conditions to contact mask substrates loaded into the substrate holder.
The design of the transfer line and all of its parts and components was done using
the computer aided design tool Solidworks. All of the design was done in house, while
the parts were machined and assembled out of house. Figs.16-17 show four external
views of the finished design as well as a photograph of the actual system installed in the
lab. The difficult portion of the transfer line design is in the details. It allows transfer of
substrates and shadow masks over distances of 13 feet, and yet the freedom of motion of
any of the mating parts is at most 1/8 of one inch. This is so that there can be minimal
strain on any given part, which will in turn minimize wear and bending of the precision
machined parts. Two good examples of this precision are in the loadlock handoff and in
the lateral fork handoff. Exemplary drawings are shown in Figs. 18-19.
b. MiST Design
-
22
-
The LP-MiST growth system, sketched in Fig. 14, integrates techniques of
physical deposition (PD) and vapor phase deposition (VPD). VPD is used for deposition
of evaporative sources, such as molecular organic materials [33], while the PD method
allows for deposition of solvated materials such as solutions of polymers, organicallydressed nanodots and colloids.
VPD of molecular organic materials has been studied extensively [33,34,35,36].
Typically, VPD occurs at reduced pressure of several Torr sustained by the constant flow
of an inert carrier gas into the growth chamber. The gas passes over hot organic sources
and rapidly carries the evaporated organic material to the substrate where it is deposited.
Compared to vacuum and spin deposition techniques, VPD gains an extra degree of
freedom by using carrier gases to transport source materials to a substrate. Indeed, in a
typical high-vacuum deposition, it is difficult to precisely control the evaporation rate of
a material by changing the temperature of the source, since the vapor pressure of the
organic materials, for example, varies rapidly with temperature. However, in a VPD
deposition, source temperatures can be held fixed while the relative concentrations of
organic constituents in the gas stream are accurately varied by adjusting the carrier gas
flow rates. For simultaneous deposition of multiple materials, their relative concentration
in the carrier gas can be precisely controlled, as demonstrated in the VPD growth of
lightly doped organic thin films used in OLEDs [34,35,36] and optically pumped organic
semiconductor lasers [34].
In addition, organic VPD has been used to deposit very
smooth, uniform amorphous films of molecules and polymers incorporated in single
heterostructure OLEDs. The absence of solvents in organic VPD growth improves the
-
23
-
polymer purity, and allows for the growth of polymers directly onto fragile molecular
organic film surfaces.
In the PD method solvated materials (QDs, polymers, colloids) are dispersed into
a fine mist of sub-10 pim-diameter liquid droplets that are injected into the growth
chamber. A stream of nitrogen gas picks up the mist and carries it to the substrate.
During the transport the solvent in the droplet evaporates due to the reduced pressure in
the chamber and elevated temperature of the gas. The initially-solvated material is now
left solvent-free in the N2 stream and is deposited on the substrate in a solvent-free form.
We expect that the benefits of PD method are similar to those of the VPD method, while
the two processes are entirely compatible as they can use the same carrier gas and similar
operating pressures. To our knowledge, PD method has never before been demonstrated,
although misting deposition of inorganic materials has been previously suggested [37].
In addition, a related technique of laser ablation of frozen solvated organics has been
demonstrated, but the technique is more complex that the PD method, and can impart
significant damage on the organic material in the form of broken bonds and bleached
color centers.
The model LP-MiST system sketched in Fig. 14 is designed to enable
co-deposition of organic materials, inorganic nano-dots, as well as soluble polymers that
cannot be evaporated due to their large molecular weight. The system consists of a
stainless steel chamber containing a hot-walled quartz (glass) cylinder.
Evaporated
organic material is delivered into the chamber by a stream of pure N2 gas, regulated by a
mass flow controller (MFC). The pressure gradient causes the organic vapor to flow
along the reactor height. In such a way, several organic materials can be simultaneously
-
24
-
introduced in the hot zone of the growth system, where the vapors mix. These are carried
to the warm zone, where the substrate is situated. Pressure in the system is maintained by
a combination of a Barotron pressure gauge and an electronically controlled throttle valve
positioned upstream of the input of the mechanical vacuum pump. The nitrogen carrier
gas is also introduced via the feeder tubes, with the gas flow adjusted using separate mass
flow controllers. Excess organic materials are condensed in a cold trap positioned at the
input to the vacuum pump.
For the PD of solvated materials (nanodots, polymers, colloids) an inlet port is
provided in center of the bottom flange. A solution mist is created by controlled injection
from an ultrasonic nozzle that is fed by a precision syringe pump capable of capable of
controlled delivery of sub-.U quantities of liquid in a continuous fashion. As the solution
mist is introduced into the chamber it is carried by the pre-heated nitrogen gas from a
separate nitrogen inlet controlled by a dedicated MFC. Exposed to the reduced pressure
and the elevated temperatures of the gas, solvent goes into the gas phase, and is pumped
out of the system. Solvated material, now freed from the solvent cage, is then delivered
to the substrate, forming a pristine film of well-characterized composition. Alternatively,
solvents can be introduced in the carrier gas to bathe the nano-dots (polymers) facilitating
chemical reaction prior and during the delivery to the substrate. Simultaneous deposition
of nanodots (polymers) and organic material should result in a uniformly doped,
homogeneous organic/inorganic mixed film, whose thickness can be precisely controlled
using an oscillating quartz thickness monitor, or an interferometric technique, and whose
doping fraction can be controlled to better than 1%.
-
25
-
VII. Future Work
Thus far, our work has concentrated on optimizing QD-LED performance for a
single sample of QDs. We must now apply what we have learned about QD-LED device
structures to create the full-color spectrum of QD-LEDs that motivated this project in the
first. While the creation of devices in the yellow, as we have done, is a nice proof of
concept, it is the red and blue LEDs that will most clearly demonstrate the utility of QDs
in OLEDs. One of the current struggles in OLED technology is the creation of a longlived, highly efficient blue device. Thus far, all of the identified blue OLED materials
exhibit instability at room temperature with time. In addition, the exciton blocking layers
that are normally necessary to allow high energy blue emission display this same
instability.
It is our hope that QDs will be the clear choice amongst blue emitting
materials, and will thus be a major focus of continuing OLED research.
In addition, we have yet to take full advantage of what the QD material set has to
offer. We have clearly identified the energy transfer mechanism as the most efficient
means of generating electroluminescence from a QD.
One clear way to control the
relative strength of this and other competing mechanisms is to vary the ZnS overcoating
thickness. QDs may be synthesized with anyway from zero to 10 monolayers of ZnS
overcoat. The QDs that we used in this work were all of 1-2 monolayers overcoating
thickness.
It is our hope that by utilizing thicker overcoatings, we might effectively
eliminate charge injection into QDs, while still allowing the exciton energy transfer to
take place. It is likely that this will yield the highest efficiency QD-LEDs achieved to
date.
-
26
-
Another variable of QDs that we have not explored is changing the organic
capping material.
The TOPO caps that we are currently using are optically and
electrically inactive. This means that they impede electrical transport in the QD:TPD
layer, and do not participate in the energy transfer mechanism. An alternative would be
to use hole conducting caps so that conduction past the QDs is aided. One could also
imagine using optically active traps, that help a cascade of energy transfer from the TPD,
to the caps, to the QDs. In this way, energy transfer could be much more complete, even
in heavily overcoated QDs.
Finally, it is our hope that as our growth system is completed, and the MiST
system becomes operational, we will have access to an ever growing number of options
in pursuing QD-LED device structures. Transparent, flexible, and stacked LEDs could
be made once we have access to a sputtering process. The MiST system will allow us to
incorporate QDs into an extremely broad range of organic host systems, regardless of
solubility compatibility. It will also allow us to place the QDs in layers besides the very
first. Finally, it will allow the deposition of QDs in a vacuum environment, adding to the
purity of the deposited layers, and thus enhancing device performance simply by further
perfecting the device material quality. All of these directions will be pursued in future
years, and it is our hope that there is much room for improvement over this current body
of work.
VIII. Conclusions
In conclusion, we demonstrate efficient LEDs utilizing QDs as lumophores in
molecular organic thin films.
These devices represent a 25-fold improvement in
luminescent power efficiency over previously reported QD-LEDs. The mechanism for
-
27
-
light emission is shown to be a combination of exciton energy transfer and charge
trapping on the QDs, with the former being desirable for maximum device efficiency. It
is clear that the limit of device performance has not yet been reached, both in quantum
efficiency and in color saturation.
Development of new deposition techniques for
generating homogeneously dispersed films of QDs in organic matrices should make
possible a much wider range of material hybrids, enabling the creation of light emitters
that are technologically competitive with state of the art organic and inorganic LEDs.
We are grateful to Universal Display Corporation and NSF MRSEC for their
support of this research.
-
28
-
References
[1]
V. Bulovid, P.E. Burrows, S.R. Forrest, Semiconductors and Semimetals 64, 255
(2000).
[2]
C. Adachi, M.A. Baldo, S.R. Forrest, S. Lamansky, M.E. Thompson, R.C.
Kwong, Appl. Phys. Lett. 78, 1622 (2001).
[3]
T. Yamasaki, K. Sumioka, T. Tsutsui, Appl. Phys. Lett. 76, 1243 (2000).
[4]
S. Moller, S.R. Forrest, J. Appl. Phys. 91, 3324 (2002).
[5]
S. Dirr, A. Bohler, S. Wiese, H. Johannes, W. Kowalsky, Jpn. J. Appl. Phys. 37,
1457 (1998).
[6]
B. D'Andrade, S. Forrest, M.E. Thompson, Adv. Mater. 14, 147 (2002).
[7]
Z. Bao, L. Chen, A. Lovinger, 0. Sapjeta, R. Jakubiak, L. Rothberg, M. Yan,
Abstr. Pap. Am. Chem. S. 219, 631 (2000).
[8]
V.L. Colvin, M.C. Schlamp, A.P. Alivisatos, Nature 370, 354 (1994).
[9]
B.O. Dabbousi, M.G. Bawendi, 0. Onitsuka, M.F. Rubner, Appl. Phys. Lett. 66,
1316 (1995).
[10]
C.W. Tang, S.A. VanSlyke, Appl. Phys. Let. 51, 913 (1987).
[11]
C.W. Tang, Appl. Phys. Let. 48, 183 (1986).
[12]
V.G. Kozlov, V. Bulovic, P.E. Burrows, S.R. Forrest, Nature 389, 362 (1997).
[13]
P. Peumans, V. Bulovic, S.R. Forrest, Appl. Phys. Let. 76, 3855 (2000).
-
29
-
G. Gu, V. Bulovic, P.E. Burrows, S.R. Forrest, M.E. Thompson, Appl. Phys. Let.
[14]
68, 2606 (1996).
[15]
G. Gu, P.E. Burrows, S. Venkatesh, S.R. Forrest, M.E. Thompson, Opt. Let. 22,
172 (1997).
G. Parthasarathy, P.E. Burrows, V. Khalfin, V.G. Kozlov, S.R. Forrest, Appl.
[16]
Phys. Let. 72, 2138 (1998).
[17]
M.A. Baldo, D.F. O'Brien, Y. You, A. Shoustikov, S. Sibley, M.E. Thompson,
S.R. Forrest, Nature 395, 151 (1998).
[18]
Y. Shirota, K. Okumoto, H. Inada, Synth. Met. 111, 387 (2000).
[19]
C.W. Tang, S.A. VanSlyke, C.H. Chen, J. Appl. Phys. 65, 3610 (1989).
[20]
C.W. Tang, Society for Information Display, InternationalSymposium, Digest of
Technical Papers27, 181 (1996).
[21]
V. Bulovid, A. Shoustikov, M.A. Baldo, E. Bose, V.G. Kozlov, M.E. Thompson,
and S.R. Forrest, Chem. Phys. Let. 287, 455 (1998).
[22]
M.A. Baldo, D.F. O'Brien, Y. You, A. Shoustikov, S. Sibley, M.E. Thompson,
and S.R. Forrest, "High efficiency phosphorescent emission from organic
electroluminescent devices," Nature 395, 151 (1998).
[23]
M.A. Hines, P. Guyot-Sionnest, J. Phys. Chem. 100, 468 (1996).
[24]
B.O. Dabbousi, J. Rodriguez-Viejo, F.V. Mikulec, J.R. Heine, H. Mattoussi, R.
Ober, K.F. Jensen, M.G. Bawendi, J. Phys. Chem. 101, 9463 (1997).
[25]
C.B. Murray, D.J. Norris, M.G. Bawendi, J. Am. Chem. Soc. 115, 8706 (1993).
-
30
-
[26]
M. C. Schlamp, X. Peng, A. P. Alivisatos. J. Appl. Phys. 82, 5837 (1997).
[27]
M. Nirmal, B.O. Dabbousi, M.G. Bawendi, J.J. Macklin, J.K. Trautman, T.D.
Harris, L.E. Brus, Nature 383, 802 (1996).
[28]
J. Liu, Y. Shi, Y. Yang, Appl. Phys. Lett. 79, 578 (2001).
[29]
R.C. Kwong, S. Sibley, T. Dubovoy, M. Baldo, S.R. Forrest, M.E. Thompson,
Chem. Mat. 11, 3709 (1999).
[30]
Ph. Tamarat, A. Maali, B. Lounis, M. Orrit, J. Phys. Chem. A 104, 1 (2000).
[31]
S.A. Empedocles, D.J. Norris, M.G. Bawendi, Phys. Rev. Lett. 77, 3873 (1996).
[32]
B.O. Dabbouis et al., J. Phys. Chem. B 101, 9463 (1997).
[33]
M. Baldo, M. Deutsch, P. Burrows, H. Gossenberger, M. Gerstenberg, V. Ban,
and S. Forrest, "Organic Vapor Phase Deposition," Adv. Mat. 10, 1505 (1998).
[34]
M.A. Baldo, V.G. Kozlov, P.E. Burrows, S.R. Forrest, V.S. Ban, B. Koene,
M.E. Thompson, "Low Pressure Organic Vapor Phase Deposition of Small
Molecular Weight Organic Light Emitting Device Structures," Appl. Phys. Lett.
71, 3033 (1997).
[35]
K.M. Vaeth and K.F. Jensen, "Chemical Vapor Deposition of Poly(P-Phenylene
Vinylene) Based Light Emitting Diodes With Low Turn-On Voltages," Appl.
Phys. Lett. 71, 2091 (1997).
[36]
K.M. Vaeth and K.F. Jensen, "Chemical Vapor Deposition of Thin Polymer Films
used in Polymer-Based Light Emitting Diodes, " Adv. Mater. 9, 490 (1997).
-31
-
[37]
V.A. Versteeg, C.T. Avedisian, and R. Raj, "Metalorganic Chemical Vapor
Deposition by Pulsed Liquid Injection," J. Am. Ceram. Soc. 78, 2763 (1995).
[38]
I.G. Hill, A. Kahn, J. Appl. Phys. 86, 4515 (1999).
-
32
-
(A)
V
Cathod
-ioooA
+
(B)
electrons and holes
form excitons
(bound e--h* pairs)
-500 A
-500 A
HTL
anode (ITO)
AO
HTL
ETL
substrate (Glass)
HO
EL Light
Figure 1
some excitons radiate
(A) Basic OLED structure. (B) Principle of operation.
1.0
0.8
C
a-NPDDCM2:Alq3
0.6
Z024
PtOEP:Alq3
A
00
0.0 -
400
500
600
700
Wavelength [nm]
800
Figure 2
Electroluminescence of Alq3 ETL,
and Alq3 doped with fluorescent DCM2 and
phosphorescent PtOEP. Chemical formula for a-NPD
HTL is also shown.
D=
120A
72A
0
-60
55A
45A
CL)
.0
33A
29A
20A
400
500
600
17A
700
Wavelength [nm]
From Murray et al, J. Am.
Chem. Soc. 115, 8706
(1993).
Figure 3
Solution absorption spectra from QDs
of various core sizes.
E
-
-
Lowest allowed
energy level
qbl&
.4
D
41
Quantum Dot Size
Figure 4
Energy band diagram demonstrating how QD
diameter contributes to higher exciton energy, and therefore
bluer absorption and emission.
8
Ahs .
U,
PL
6
000
TPD:QD
C6-
3000
E 4-
.4
2000
C
CU)
C2
Cn
QD
2-
300
Figure 5
-1000
350
400
450 500 550
Wavelength (nm)
600
650
Absorption and photoluminescnece spectra
of QD neat films and TPD:QD mixed films. Calculations
utilizing this data reveal that our mixed films are roughly 5%
by volume QD.
Ak
Vacuum Level = 0 eV
EA=2.1
TPD
2,)
EA=3.1
Alq3
c
2
2*
W
EF=4-7.
-4.7
ITO
EA=.4..3
EA=4.
S1=5.4
K
E=3.7
Mg
-
IE=5.8
IE=6.8
Proposed energy level diagram of device I.
Figure 6
Where possible, values for ionization energy (IE) are taken
from photoemission spectroscopy measurements [38].
Electron affinity levels (EA) were then calculated using
optical absorption data. QD levels shown are from calculated
values [26].
Device
caps
O_5[DJ
shell
c
A
A
TOPO
CdSe
Score
Device III
Device II
I
m
F
-..
4Onm Alq
4ps
40nm TPD:D"
4Onm AlQ
5nm TP
40nm TPD:QD
fo=0.36
Glass
fQD
fl-=
A
2
0 38
0q3
.
400 450 500 550 600 650
fQD=0.85
Glass
ILGlass
A3
fAlq3=
;
400 450 500 550 600 650
RRA
4Onm Alc!
10nm TAZ
nm TPD:QD
fT
D006
009
EL
400 450 500 550 600 650
Wavelength [nm]
Figure 7
Electroluminescence spectra and structures for three QD-LEDs. Dashed lines show the
decomposition of spectra into Alq3 and QD components. The inset shows a cartoon of a core-shell type QD.
o1-
Luminance of Device I [Cd/M 2]
1
10
100
1000
.) 0.8 -
0.6 C
0.4 E02
Device I I
3 0.2 0
cD0.1 I-
w
Figure 8
1
10
100
0.1
2
Current density [mA/cm ]
External quantum efficiency versus current
density for the three devices shown in Fig. 7. The peak
efficiency of device I is q-=0.61% at 7mAlcm 2 and 6.3 V, which
corresponds to a luminance of 120cd/m 2.
CE
102.
E lo
E 10'
Device 11
Device Ill
C
Device I
Q)
-'-106
0 .1
0.5
1
5
10
Voltage [V]
Current-voltagedata for the three devices
Figure9
shown in Fig.7. In all cases I xV", where n is between 8 and
10.
1,11,111
1,11
Excitons created
on Alq3
Excitons created
on TPD
Semission],
1,.
TPD emission
III
4
1,111
Excitons created
on QD
emei
,1
Alq3 emission
I'll
Flow-chart of excitonic processes in QDFigure 10
LEDs. Roman numerals refer to the devices depicted in Fig.7.
111
QD emission
1,11,111
. . ..
............................
.
.. .. ..
ZnEt 2
ZnS (TMS) 2S
4-
TOP/TOPO
-200 0 C
/ j\\\
(b)
10001000:
U)
C
a)
Z
100-
100-
C
1010400
500
600
700
800
Wavelength [nm]
900
1000
400
500
600
760
800
90
1000
Wavelength [nm]
Figure 11
Electroluminescence spectra as a function of current of QD-LEDs with the
structure of device I from Fig. 7 and containing (a) ZnS coated CdSe nanodots and (b) uncoated
CdSe nanodots. Current increases from 0.1 to 2 mA from the bottom to the top of both plots
50-
1.0-
(a)
45
40-
E
C
(b)
0.8-
35C7
U)
30-
-0 0.6-
25
20-
0.40
15-
C
10-
C
5400
0.2-
C
450
500
550
600
Wavelength [nm]
650
700
400
450
500
550
600
Wavelength [nm]
650
Electroluminescence spectra as a function of current of QD-LED of device I in
Figure 12
Fig. 7. In (a) plot is unnormalized and in (b) it is normalized to the 570nm peak. This shows the
current dependence of fAlq3 and fQp.
700
Chamber
Ante
Station
Probe
Probe Station
with Cryostat
AFM
STM
Eva porative Deposition
Shado w
Mask Storage
Ante Chamber
and Oven
Wet N2 Glove Box
Dry N2 Gtove Box
Chemic at Vapor Dep.
-
--
*o*
Sputtering
Load Lock
with Sampte Storage
Laminar FRow Hood
-------------
Physical &t
Vapor Phase Dep.
Integrated materials and devices growth system consisting of the materials
Figure 13
growth chambers based on physical and vapor phase deposition, sputtering, chemical vapor
deposition, and evaporative deposition with in-situ shadow masking. Samples are analyzed in
the analytical and AFMISTM chambers, and all connected to the central transfer line with load
lock and an integrated, nitrogen-filled glove box.
1
to pump
thickness
monitor
transfer to the
rotating
substrate
DI
Integrated
I
to pump
-
holder
-
--
*|
Growth System
shutter
lamp-
1
&%4 1d
Sublimed Molecular
Organic Materials
(in MFC controlled
nitrogen carrier gas)
heated tube
I
of
sourc
"MMURia
I
-
lamp
glass shield
ultrasonic nozzle
Nitrogen in (from MFC)
= Solution in
(liquid dispensing
is controlled by a
syringe pump)
Proposed Misted Solution
Fig. 14
Transport (LP-MiST) deposition chamber.
Figure 15
Solidworks drawing of the substrate holder
and shadow mask holder self-assembled stack. This example
holds Y" substrates, our most typical substrate size.
Solidworks drawings of the transfer line in
Figure 16
various orientations. Two photos showing the actual 12' line
in our lab are also included.
Figure 17
Solidworks endview drawing of the transfer
line. The photograph is of the system from the same
orientation. Note the rails and the shadow mask holder stack
that allows for in-situ masking of substrates.
Figure 18
Photograph of the caboose that rides the
rails of the 13' chamber, and transports the substrate holders
and shadow mask holders. Solidworks view of both the engine
and the caboose in their "hand-off' configuration.
aQZ-i,
--
F
1
o
o
t
0
0
0
O0
0
0
mechanics that ensure perfect insertion of the transfer fork
into the substrate holder. Solidworks perspective view of the
same.