Alternative Methods and Materials for Patterning
Organic Thin Film Electronics
by
Matthias Erhard Bahlke
B.A. Physics (2009)
Bard College
B.S. Electrical Engineering (2009)
Columbia University
M.S. Electrical Engineering and Computer Science (2011)
Massachusetts Institute of Technology
Submitted to the Department of Electrical Engineering and Computer
Science
in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 2014
c Massachusetts Institute of Technology 2014. All rights reserved.
Author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Department of Electrical Engineering and Computer Science
May 14, 2014
Certified by . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Marc A. Baldo
Professor of Electrical Engineering
Thesis Supervisor
Accepted by . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Leslie A. Kolodziejski
Chair, Department Committee on Graduate Students
2
Alternative Methods and Materials for Patterning Organic
Thin Film Electronics
by
Matthias Erhard Bahlke
Submitted to the Department of Electrical Engineering and Computer Science
on May 14, 2014, in partial fulfillment of the
requirements for the degree of
Doctor of Philosophy
Abstract
Photolithography’s accuracy and scalability have made it the method for sub-micronscale definition of single-crystal semiconductor devices for over half a century. The
ultimate goal for OLED manufacturing, however, is to replicate the widespread success of photoresist lithography without the use of the types of resists, solvents, and
etchants traditionally used—organic small molecules are simply not compatible with
these tools. Hence, there is motivation for a renewed examination of variants of this
inherently parallel, high speed approach. This work investigates the use of chemically
inert resists that rely on clearance mechanisms not found in traditional lithography.
These primarily include employing phase changes for lift-off patterning thin films of
organic semiconductors and metals, and also propose and discuss the use of combustible and magnetic materials.
Thesis Supervisor: Marc A. Baldo
Title: Professor of Electrical Engineering
3
4
Acknowledgments
First and foremost I’d like to thank my research advisor, Marc, for his ideas, interest
and encouragement—I’d have been unable to complete this work without his help. I’d
also like to thank Hiroshi Mendoza for his sustained optimism and undying motivation
and work ethic. Phil Reusswig is always thinking of new ideas and would lead off
even his greatest ground-breaking ideas with “this is probably not very good, but...”
He’s been an amazing colleague, roommate, and, most of all, friend. I learned so
much about modern OLED materials, devices and their characterization, bicycles
and the best bread in the world from Sebastian Reineke. Working with him and
Markus was always exciting even when results were less than perfect. Mingjuan Su
and I put a great deal of work into some interesting material that did not fit into the
scope of this thesis, but she’s a brilliant and enthusiastic synthetic chemist and friend.
Nicholas Thompson has been very helpful with his continued interest, feedback and
suggestions. It will be very exciting to see televisions with the materials and devices
Nick and Mingjuan develop at UDC on the shelves of stores and someday in our
homes. Apoorva Murarka saved us a lot of time donating one of his old MEMS
stamps and assisting with making the first of my own. Jiye Lee was an amazingly
self-motivated driving force in the group and was the person that got me started in
the group. Priya Jadhav provided necessary and much-appreciated assistance with
experimental design. Jason Sussman always has new ideas to discuss and has a great
historical knowledge of the physicists that laid the foundation we built on. Many
thanks to Jon Mapel for his assistance with scaling and market evaluation. I’m very
grateful to Cathy Bourgeois for being on top of everything and helping out every
step of the way. I’d like to acknowledge the rest of The Soft Semiconductor Group
that I’ve worked along side of: Paul Azunre, Daniel Congreve, Jean Anne Currivan,
Sumit Dutta, Markus Einzinger, Tim Heidel, Brian Modtland, Carlijn Mulder, David
Ciudad Rı́o-Pérez, Carmel Rotschild, Saimi Siddiqui, Amador Velázquez, Mengfei
Wu, Tony Wu, and Allen Yin for assistance and helpful discussions along the way.
I’m indebted to John Kymissis and the Columbia Laboratory for Unconventional
Electronics for helping me realize that graduate school in engineering was right for me.
Marshall Cox, Eddy Hsu, Zhang Jia, Vincent Lee, Nadia Pervez, Samuel Sabbarao,
and John Sarik are amazing teachers, friends and colleagues. Jon Beck and I worked
on projects together as undergraduates and, even though I left CU for MIT, we
worked concurrently on our PhD’s sharing ideas, advice, and feedback. Security guru
Michael Halsall held a continued interest in my work and inadvertently sprouted my
great interest in his. I appreciate Leslie’s patience in the final months of this work
when my free time dwindled and my mind was never off the thesis and projects.
Outside of those that assisted with theory and experimental work, I’d like to thank
my mother and father for their support as well as Marlene, George, Sarah and Cora.
5
6
Contents
1 Introduction
1.1
1.2
17
Organic Light-Emitting Diode Displays . . . . . . . . . . . . . . . . .
18
1.1.1
Current Commercial Status . . . . . . . . . . . . . . . . . . .
19
1.1.2
Structure of an OLED . . . . . . . . . . . . . . . . . . . . . .
20
Motivations for an Alternative Patterning Process . . . . . . . . . . .
21
1.2.1
Fine-Metal Masks . . . . . . . . . . . . . . . . . . . . . . . . .
22
1.2.2
Inkjet Printing . . . . . . . . . . . . . . . . . . . . . . . . . .
23
1.2.3
Photolithographic Processes Involving Solvents . . . . . . . . .
25
1.2.4
Other Ways to Pattern Organic Thin Films . . . . . . . . . .
25
2 Physics of Organic Materials and Devices
27
2.1
Electrons, Excitation, and Spin . . . . . . . . . . . . . . . . . . . . .
27
2.2
The OLED: Turning Charge into Light . . . . . . . . . . . . . . . . .
29
3 The Sublimable Mask Lithographic Process
33
3.1
Theory of Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . .
34
3.2
Process Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
35
4 Sublimable Mask Patterning
39
4.1
Considerations for Demonstration . . . . . . . . . . . . . . . . . . . .
39
4.2
Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . .
40
4.2.1
Patterning the Resist with Resistive Heating . . . . . . . . . .
44
4.2.2
Patterning the Resist with a Stamp . . . . . . . . . . . . . . .
51
7
4.3
OLEDs in a Cold Environment . . . . . . . . . . . . . . . . . . . . .
62
4.4
Patterning the Resist Photolithographically . . . . . . . . . . . . . .
64
5 Future Directions
67
5.1
Combustion Lithography . . . . . . . . . . . . . . . . . . . . . . . . .
67
5.2
Magnetic Resists . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
68
6 Conclusion
71
A Selected Molecules Used in this Work
85
B Triple Points of Some Gases
89
C MATLAB Code
91
8
List of Figures
1-1 Dimensions of subpixels making up each of the over 8 million pixels
found in a 55-inch UHDTV. The area obstructing the top of the pixel
is from the driving transistors. It is worth noting that this is an oversimplification as the variation in performance of materials for different
colors in an OLED display necessitates different sizes for each color. .
19
1-2 A fine-metal mask used for arrays of OLED subpixels[131] (top) and
some of the masks we use in our group(bottom). . . . . . . . . . . . .
22
1-3 Inkjet printing deposits droplets of dissolved organic compounds that
leave a solid thin film after the solvent evaporates. . . . . . . . . . . .
24
1-4 Kateeva’s YIELDjetTM Gen-8 inkjet manufacturing equipment[62]. . .
25
2-1 Figure from [5] showing results of density function theory(DFT) calculations of pentacene and fullerene C60 exciton and charge-transfer
states. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
28
2-2 Organic transport level diagram showing the electroluminescence process. The hole and electron are injected from their corresponding electrode and move through their transport layers to meet and form an
exciton. The exciton recombines dissipating its energy through a photon.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
2-3 (a) Thermally activated delayed fluorescence (TADF) energy diagram.
(b) Some of the molecules that Uoyama et al. presented in their seminal TADF paper that this figure is borrowed from (Reprinted with
permission from [119]. Copyright 2012 Nature Publishing Group.). . .
9
31
3-1 An SEM image of The Harvard Nanopore Group’s demonstrated patterning of chromium using water ice resist. Reprinted with permission
from [65]. Copyright 2005 American Chemical Society. . . . . . . . .
34
3-2 Inside cover of December 2012 issue of Advanced Materials featuring
sublimation lithography[7] . . . . . . . . . . . . . . . . . . . . . . . .
35
3-3 General phase diagram. The arrow shows the ideal region of operation
with respect to sublimation lithography. Deposition of the mask takes
place along the low temperature region of the arrow and sublimation
and lift-off take place along the higher temperature region of the arrow. 36
3-4 Phase diagram reflecting full process flow parameters. The defining of
the mask pattern is performed immediately prior to the deposition of
the thin film. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
37
3-5 Simplified process flow for sublimation lithography. (not to scale) (a)
Begin with a cooled substrate to facilitate resist deposition. (b) Deposit
resist. (c) Selectively pattern resist. (d) Deposit desired thin film. (e)
Lift-off resist leaving patterned thin film. (f) Repeat as necessary to
complete device. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
37
4-1 The side of the evaporation chamber showing the feedthroughs and
liquid nitrogen reservoir. . . . . . . . . . . . . . . . . . . . . . . . . .
41
4-2 Flange making up the bottom of the liquid nitrogen reservoir. The
substrate holding plate is bolted in with a glass sample tightly secured.
The washer-mounted thermocouple can be seen towards the top of the
image mounted with a bolt. . . . . . . . . . . . . . . . . . . . . . . .
42
4-3 Phase diagrams of CO2 and H2 O extrapolated (broken lines) from data
(solid lines) in references [83] and [120]. The “×” represents the process
operating point of 77 K at 106 Torr. The curves are extrapolated using
the Clapeyron equation[112]. Figure from [2] . . . . . . . . . . . . . .
43
4-4 The pattern of the ITO used in resistive heating experiments. The
black regions are ITO; the white regions are bare glass. . . . . . . . .
10
46
4-5 A mask of CO2 patterned for a 100 µm-wide line. The substrate holder
is rigged with probe-tipped clips for electrical contact. . . . . . . . . .
47
4-6 In situ photograph and schematic representation of the resistive heating method of patterning the frozen mask. Electrical contacts on the
substrate allow current to be driven at a current density of 625 kA cm2
through a 160 nm-thick strip of ITO on the substrate. The red arrows
in the schematic drawing indicate the flow of current through the ITO
enabled by affixed copper foil tape[2].
. . . . . . . . . . . . . . . . .
48
4-7 An exmaple of photoluminescence from a 100 µm-wide line of DCJTBdoped Alq3 (top), and photoluminescence(bottom-left) and a micrograph(bottomright) from a 100 µm-wide line of Alq3 patterned by resistive heating
of ITO. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
49
4-8 A micrograph from a 100 µm-wide line of silver patterned by resistive
heating of ITO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
4-9 Simplified process flow for sublimation lithography using a frozen CO2
resist (not to scale). (a) Cool substrate below 100 K. (b) Freeze on
thin film of CO2 . (c,d) Pattern CO2 film by heating selectively. Here
a stamp is pressed against the resist to remove some areas. (e) Deposit desired organic or metal thin film by thermal evaporation. (f)
Warm substrate to sublime CO2 , thus lifting off unwanted material.
(g) Repeat steps (a-f) as necessary to complete the device[2]. . . . . .
51
4-10 2D simulation of thermal diffusivity in the stamping process. . . . . .
53
4-11 A micrograph of the 115 µm-tall pillar SU-8 stamp(top) used to pattern
circles of Alq3 (bottom) . . . . . . . . . . . . . . . . . . . . . . . . . .
11
54
4-12 Schematic of the updated experimental stamping setup showing the
process-critical components as described in the main text. The entire setup is located within the standard thermal evaporation chamber
(Angstrom Engineering). The arrows by the stages indicate their respective direction and range of motion. Both the stamp and the microscope used for positioning and stamping evaluation are located on
the smaller stage as indicated[2].
. . . . . . . . . . . . . . . . . . . .
55
4-13 Micrograph of a region of the SU-8 stamp used to selectively sublime
regions of the CO2 resist for 78 µm-pitch patterns. In the center of the
image is a pillar that has been knocked over displaying the profile of
the pillar[2]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
57
4-14 Optical micrograph (a) of a 78 µm pitch-patterned mask of CO2 defined
via contact with a stamp. The corresponding inset (b) shows the Alq3
thin film after deposition and lift-off. Photoluminescence micrograph
(c) from the same film[2]. A longpass filter is used to lessen pump
light’s presence in the image. . . . . . . . . . . . . . . . . . . . . . . .
58
4-15 A false-color topography obtained by an optical interferometer (top)
and a cross section of the same (bottom) detailing the profile of the
patterned organic pixels of figure 4-14[2]. . . . . . . . . . . . . . . . .
59
4-16 Examples of red and green subpixels side-by-side. The left two images
show actual red and green subpixels next to each other whereas the
right image exemplifies some of the complications that arose in the
attempts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
61
4-17 Photoluminescence micrograph of a free-floating 50 nm-thick film of
Alq3 folding away from the substrate yet remaining intact. It measures
about 150 µm on a side. . . . . . . . . . . . . . . . . . . . . . . . . .
62
4-18 External quantum efficiency versus current density of OLEDs grown at
T = 112 ± 24 K and room temperature (a). The normalized electroluminescence spectrum is indistinguishable from the room temperature
control device (b) and device thin film stack (c) are also shown[2] . .
12
64
5-1 SEM micrograph of nitrocellulose on silicon patterned via an argon
ion beam by Geis et al. in 1983. Reprinted with permission from [41].
Copyright 1983 American Vacuum Society. . . . . . . . . . . . . . . .
68
5-2 Suggested process flow for using magnetic materials as dry lift-off resists. (not to scale) (a) Begin with clean substrate (b) Apply magnetic
resist where subsequent thin film is not desired (c) Deposit thin film
(d) Remove resist and unwanted thin film areas with a magnet or other
strong magnetic field (e) Repeat as necessary to complete all layers of
desired device. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
69
5-3 Jacobs et al. stamped elecrolets on PMMA to position the above materials. Reprinted with permission from [58]. Copyright 2001 American
Association for the Advancement of Science. . . . . . . . . . . . . . .
70
5-4 Magnet domain arrangement of a flexible vinyl magnet and the resulting field lines. Image from [123] . . . . . . . . . . . . . . . . . . . . .
70
A-1 Aluminium tris(quinolin-8-olate) (Alq3 )[125] . . . . . . . . . . . . . .
85
A-2 Carbon dioxide (CO2 ) . . . . . . . . . . . . . . . . . . . . . . . . . .
85
A-3 4-(Dicyanomethylene)-2-tert-butyl-6-(1,1,7,7-tetramethyljulolidin-4-ylvinyl)-4H-pyran (DCJTB)[125] . . . . . . . . . . . . . . . . . . . . . .
86
A-4 SU-8 Photoresist[125] . . . . . . . . . . . . . . . . . . . . . . . . . . .
86
A-5 Tris[2-phenylpyridinato-C2 ,N ]iridium(III)(Ir(ppy)3 )[11] . . . . . . . .
87
13
14
List of Tables
1.1
Typical display specifications of current competing mobile displays[3][4][106][55]. 18
15
16
Chapter 1
Introduction
O
rganic semiconductors devices are thin and efficient, but have yet to live
up to their full potential as the basis of cheap optoelectronic devices. The
problem is not one of material costs, but of obstacles to patterning for
large-scale fabrication. This work provides an alternative manufacturing technique
employing phase-change resists that may allow for the retirement of the fine-metalmasking methods currently used.
This chapter introduces the physics, working principles, commercial status and the
as-of-yet untapped potential of organic light-emitting diodes. It also explains how a
dry lithographic process could overcome present challenges in scaling up production
of OLEDs on larger sizes of mother glass.
Chapter 2 provides a brief introduction of organic materials and the physics that
enable OLEDs. Modern emitter concepts are discussed.
Chapter 3 introduces the sublimable mask patterning idea and provides a stepby-step overview of the process.
Chapter 4 describes the approach and design of a proof-of-concept implementation
using carbon dioxide ice as the resist.
Chapter 5 discusses future directions of dry patterning of organic thin films.
Chapter 6 summarizes the results and describes options for large-scale patterning
of the resists and the under- and overlying thin films. Limitations of techniques and
the feasibility of such processes are briefly discussed.
17
Table 1.1:
Typical
displays[3][4][106][55].
display
specifications
of
current
competing
mobile
Device
Display Size (in) Resolution Technology ppi1
Apple iPhone 5S Retina
4
1136x640
IPS LCD
326
Apple iPad Retina Mini
7.9
2048x1536
IPS LCD
326
Samsung Galaxy S5
5.1
1920x1080
AMOLED
432
HTC One M8
4.7
1920x1080
LCD
441
1.1
Organic Light-Emitting Diode Displays
A 32-inch full high-definition television (HDTV) requires a pixel size of a 370 µm-per
side square further broken up into 3 subpixels: one for red, one for green and one
for blue to reproduce the necessary colors required for full color. Having features
with dimensions of about 120 µm necessitates patterning accuracies on the order of
10 µm or less. For larger displays of the same resolution the pixel dimensions are
larger, but for smaller displays the constraints are tighter. The move to even higher
resolutions brings these patterning accuracy requirements tighter still. Ultra high
definition resolution in a 55-inch display, for example, means pixels 320 µm on a side
with subpixel widths of around 100 µm as shown in figure 1-1.
The highest resolution displays are found in mobile electronics like smartphones.
They must reproduce great amounts information clearly on a screen of about 4 to 5
inches. The size, resolutions and pixel-per-inch (ppi) specifications of some of today’s
best displays are shown in table 1.1. 450 ppi corresponds to a pixel pitch of about
56 µm; 350 ppi corresponds to a pixel pitch of about 73 µm; 250 ppi is more like 101
µm. OLED manufacturers are at present unable to realize pixel pitches much less
than 60 µm due to the problems laid out in section 1.2.
1
I’d like to note that beyond ∼326 ppi the human eye can no longer discern individual
pixels[111]—hence the name “Retina Display”
18
320μm
RGB
100μm
320μm
Figure 1-1: Dimensions of subpixels making up each of the over 8 million pixels found
in a 55-inch UHDTV. The area obstructing the top of the pixel is from the driving
transistors. It is worth noting that this is an oversimplification as the variation in
performance of materials for different colors in an OLED display necessitates different
sizes for each color.
1.1.1
Current Commercial Status
Up until 2013 OLED televisions were only available in 11-inch and 15-inch sizes. The
resolutions of these displays were sub-full high-definition. Following these, 55” fullHD units were introduced first by LG Electronics Display and soon after Samsung,
although they came at a premium; a customer taking one of these home would be
very lucky to pay less than $10,000. As production yields and fabrication techniques
have improved these prices have slowly begun to creep closer to about half of their
introductory prices. Panels as large as 77 inches have been demoed at Ultra High
Definition resolutions by LG[29] and Panasonic has shown its own 56-inch 4K televisions using inkjet printing rather than thermal evaporation for its material deposition
and patterning[93].
Samsung is the OLED industry leader and opened a new $2.1 billion 5.5thgeneration OLED factory in the summer of 2011 to produce larger active-matrix
OLED HDTVs[20][74] and has a 6th-generation fabrication facility on order[13]. Samsung’s yearly investment in OLED technology in 2011 was about $4.8 billion, matching
its investment in liquid crystal display technology[75].
19
In 2010 Samsung Mobile Display announced that after years of investment, the
OLED division became profitable[64]. It is of interest here to note that the market
contribution from televisions is negligible at this time, so there is much room for
growth. Reducing the cost of OLED displays by solving yield and scaling complications is key. The current organic electronics market stands at an estimated $11-12
billion[31][107] and is expected to increase to $44 billion by 2019[91]. It is expected
that by 2015, OLED HDTVs alone will be a $2 billion industry[23][51].
1.1.2
Structure of an OLED
OLEDs have a simple general structure and principle of operation. The first OLED
demonstrated by C.W. Tang and S.A. VanSlyke in 1987[117] consisted of a monopolar
transport layer and an emissive layer sandwiched between a transparent anode and
an alloyed metal cathode. Today’s OLEDs are a little more complicated: they make
use of both singlet and triplet excitons for light generation[8, 34, 88, 101, 119, 124,
132, 135], quadrupling efficiencies relative to those devices relying on fluorescence
alone[9]. Modern devices consist of an anode, hole-injection and hole-transport layers (HIL and HTL, respectively), host and dopant materials, electron-transport and
electron-injection layers (ETL and EIL, respectively), blocking layers and a cathode.
The anode or cathode must be transparent to allow light emission, or both for a
transparent display[44, 71, 76, 95, 97, 128, 136].
The reason for so many layers is to boost efficiency by increasing the probability
of charges recombining and, in turn, generating light. When an external voltage is
applied across the layers of an OLED, holes are injected from the anode and electrons
are injected from the cathode. The positive and negative charges meet at a charge
recombination zone (the emission layer) in the middle where light emission occurs.
This is covered in greater detail in chapter 2.
20
1.2
Motivations for an Alternative Patterning Process
Silicon is undoubtedly the most widespread semiconductor and technology would be
nowhere close to where it is now without it. It’s manufacturability[92, 127], mobility,
and cost have enabled everything from small photodetectors, thermal diodes, and
microelectricalmechanical systems (MEMS) to processors for supercomputer nodes
containing billions of transistors. However silicon is mostly limited to rigid integrated
circuits and small devices. The class of high-performance crystalline semiconductors
can make for highly efficient LEDs; to make an LED TV, on the other hand, costs
would be prohibitively high in the case of individually wired devices, not to mention
a single crystal substrate a square meter in area with millions of pixels.
Organic materials, on the other hand, are generally less expensive, but have the
problem that most are sensitive to solvents. Every step of the lithography process
must be completely dry once any organic thin films are introduced2 . This means that
a preliminary solution-processed step is allowable, but if you are trying to pattern
multiple successive layers (like in the case of separate red, green, and blue subpixels),
it can’t be a defining part of the patterning process. In the case of color-by-white, as
developed by Kodak and Sanyo[122, 130] and used by LG[80] and eMagin[33], only
one patterning step is necessary if at all.
The only OLED displays that have reached profitability are in the small and
mobile display markets. While larger displays are available, quantities and choice of
products are limited and prohibitively expensive. The prices of larger displays would
come down if the generation of display production (which corresponds directly to the
size of the mother glass that the displays are grown on) were able to increase. This
is currently limited by manufacturing processes.
The main competitor to OLEDs is liquid crystal displays (LCDs)(which are made
in production plants rated about five generations higher than those of OLEDs). While
an inherently more complicated design (involving polarizers, color filters, spacers, and
2
With the exception of the subset of materials specifically formulated for solution processing
21
liquid reservoirs[90]), their photolithography-compatible materials allow for simple
scaling up of pixel electrode, transistor and color filter definition, thereby allowing
larger mother glass production and consequently greater throughput[22, 63, 82]. The
section that follows describes the current methods used to pattern organic electronic
thin films.
1.2.1
Fine-Metal Masks
Current OLED patterning technology uses thin sheets of steel (see figure 1-2) that
are used to define pixels during the thermal evaporation of organic semiconductors
and metals. Placing these fine-metal masks (FMM) across sheets of glass like stencils
to produce features on the order of 10 µm introduces complications.
Figure 1-2: A fine-metal mask used for arrays of OLED subpixels[131] (top) and some
of the masks we use in our group(bottom).
22
FMMs are fragile and, due to the nature of the masking process, must be cleaned
after a number of growths to prevent defects caused by debris. This puts stress on
the features of the mask and can result in tears and bending. They cost on the order
of $200,000 and need to be replaced every one to two months. Moreover, temperature
deviations during the deposition can cause thermal expansion and contraction limiting the feature size[114]. Because the mask is not projected, but instead renders a
1:1 reproduction, incorporating multiple masks in a fabrication line requires tedious
alignment[39]. This puts a bottleneck on throughput that, if eliminated or improved
upon, would greatly reduce costs.
One temporary measure introduced to reduce the complications of a FMM that
spans an entire mother glass is to step a smaller area mask across for multiple material
depositions. This is termed “Small Mask Scanning technology” by Samsung[108]. The
mask-area-limiting factor becomes only the size of an individual display, but the takt
time is increased by at least the number of displays per mother glass over a mother
glass-spanning mask. In addition to this increased takt time, the accumulated layers
would cause a mask to wear out sooner due to the higher exposure of incident metal
and semiconductor sublimation per mother glass.
1.2.2
Inkjet Printing
Inkjet printing is another potential method for full-scale fabrication of organic electronics. The oft-touted benefits include production without vacuum equipment, low
operating costs and high material-use efficiency[52]. An inket head operating on an
XY-stage may also allow for simple pattern programming.
Unfortunately, inkjet printing has its own limitations. Such an inherently serial process requires an enormous number of printing heads to allow for reasonable
throughput[10, 113]. Even then, the organics in such a system must usually be put
in solution before being printed requiring the replacement of many industry-leading
small molecule organics with molecules with poorer performance3 . Printing solutions
3
It should be noted that this material limitation is not the case for molecular jet printing technologies in which molecules are evaporated from a micrometer-scale printing head[19].
23
Figure 1-3: Inkjet printing deposits droplets of dissolved organic compounds that
leave a solid thin film after the solvent evaporates.
also have issues with drop uniformity due to surface tension during the evaporation
of the solvent[96][27][39], but this is actively being researched.
In late 2013, Kateeva announced their YIELDjetTM inkjet manufacturing equipment specifically for OLED processing.[62] YIELDjetTM is the culmination of years
of research focused on material and environmental purity, film uniformity, and device
and processing reliability. The intention is compatibilty with Gen-8 manufacturing
lines and Kateeva claims that their current Gen-8 unit, shown in figure 1-4, is capable
of covering a full mother glass with appropriately patterned organics in five minutes’
time. At the time of writing, no consumer products have been released, but the
technology shows much promise.
24
Figure 1-4: Kateeva’s YIELDjetTM Gen-8 inkjet manufacturing equipment[62].
1.2.3
Photolithographic Processes Involving Solvents
Photolithography is widely used for patterning in the semiconductor industry, but is
difficult to apply to the production of organic semiconductor devices. While solutionprocessable polymer devices have been photolithographically patterned[72, 133, 134],
the resists, solvents and etchants are chemically incompatible with all small-molecule
organic semiconductors; in our group’s experience, some common compounds (Alq3 ,
TPBi, and BPhen to name a few) have proven soluble in the fluorous solvents used in
orthogonal photolithography employing hydrofluroethers. More general photolithographic processes have required an intermediary polymer barrier [28] or the use of
super-critical CO2 [56]. The work proposed here investigates the use of a few dry
chemically inert resists that can be deposited and patterned in situ.
1.2.4
Other Ways to Pattern Organic Thin Films
Many other patterning methods have been developed and tested in universities and
industry. Some, like in the case of laser-induced thermal imaging (LITI)[61, 73, 77,
78], have been seriously considered for use in full production lines. For more details
25
and more general information on patterning and manufacturing OLEDs, I recommend
Martin B. Wolk’s chapter Patterning of OLED Device Materials in [60].
26
Chapter 2
Physics of Organic Materials and
Devices
mall molecules are the main players in this work: relatively low molecular
S
weight compounds consisting mostly of carbon and hydrogen with nitrogen,
oxygen, sulfur, and/or some heavy metal or metals. Because of the strength
of the intermolecular bonds, they tend to form amorphous morphologies as opposed
to crystalline ones. This weak bonding also results in the electronic and photophysical
properties for a solid being similar to those of the single molecule. This also results
in low evaporation points. These two characteristics allow one to rather easily make
heterojunctions of multiple different materials that exhibit properties of the individual
materials.
2.1
Electrons, Excitation, and Spin
Electrons are not arranged the same way on molecules as they are in atoms because
molecules are made up of arrangements of multiple atoms of various elements in
relatively complicated geometries. Instead, electrons in a molecule are considered to
be in molecule orbitals[85] that describe the locations around a molecule that help one
determine overlap of orbitals between adjacent molecules, approximate energy levels
of a molecule, and determine what kinds of interactions are possible. There are a
27
number of mathematical formulations that allow predictions of values and properties
of a molecule like density functional theory (DFT)[54, 68, 94] and the Hartree-Fock
methods[35, 36, 50, 115] for example, the details of which are outside the scope of
this work and make up an entire field of research on its own. While singlet fission and
charge dissociation are not the focus of this work, figure 2-1 exemplifies the value of
these calculations.
Figure 2-1: Figure from [5] showing results of density function theory(DFT) calculations of pentacene and fullerene C60 exciton and charge-transfer states.
When an incident photon of adequate energy is absorbed by an organic molecule,
an electron is excited creating an electron-hole pair called an exciton. This is in
contrast to what happens in common inorganic semiconducting materials like silicon
where you’d almost immediately have a free charge. This is because of the relative
permittivity r being so much lower in organics. The high dielectric constant of
inorganic materials makes for binding energies so small that room temperature phonon
vibrations are sufficient to dissociate the excitons formed therein. The reverse of
dissociation is also possible, in which case an electron and hole combine to form an
exciton.
28
Excitons
The wavefunction describing two electrons must be antisymmetric. Thus, if the spins
of those two states are symmetric, the spatial factor of the wavefunction must be
antisymmetric. The greater distance associated with this spatial portion means the
electons shield eachother less and are as a result lower energy, more tightly-bound
states. These are known as triplet excitons as there exist three unique possible spin
pairings as opposed to singlet excitons that have only one[98]. Singlet radiative
relaxation is known as fluorescence whereas triplets quantum-mechanically forbidden
relaxation is known as phosphorescence[79, 98]. These states importance in OLEDs
is discussed further in the next section.
When considering a molecule’s interaction within a device or system, the two
orbitals at the cusp of the ground state and the excited state are considered: the
highest occupied molecule orbital (HOMO) and lowest unoccupied molecular orbital
(LUMO). They serve as a guide to electron and hole transport1 in a molecule and
device as well as the excitation energy of the molecule. The difference between the
HOMO and LUMO is often referred to as the band gap and this level is the sum of
the optical gap and binding energy (which, itself, is made up of the electron Coulomb
interaction and exchange energy)[67, 98].
2.2
The OLED: Turning Charge into Light
As touched on in the introduction, an OLED is a suitably arranged stack of metallic
and organic thin films enabling the injection of charges into an organic layer that
later radiatively relax, i.e., photons are emitted from these induced excited states.
Which particular layers are used is determined by examining the HOMO and LUMO
of the emissive material. In order to get electrons and holes into this material, it
is necessary to use charge transport materials with a low enough LUMO and a high
enough HOMO (both in terms of energy), respectively, paired with appropriate work1
The HOMO and LUMO levels are for this reason sometimes referred to as organic transport
levels.
29
function electrodes for injection of each corresponding type of carrier[8, 117]. It isn’t
as simple as that, though, as a low-energy transport material might serve as a loss
mechanism if excitons would find lower energy states in it or if the transport material
isn’t transparent to the EML’s emitted light. It is for this reason that wide-gap
blocking layers are sometimes used[9] .
In essence, it is important to confine excited states in the emissive material for
an OLED to be efficient[117]. Figure 2-2 shows a generalized energy level diagram of
an OLED with the path of each carrier from electrode to emission. The electron and
hole transport layers are often abbreviated as ETL and HTL, respectively, and most
modern emitter materials are grown dilute in a wide gap host to prevent annihilation
effects stemming from aggregation[102, 103, 104]. The emissive layer, being either a
single material or an emitter doped into a wide-gap host, is abbreviated as the EML.
Energy
Host
HTL
Anode
+
*
_
EML ETL Cathode
hv
Figure 2-2: Organic transport level diagram showing the electroluminescence process. The hole and electron are injected from their corresponding electrode and move
through their transport layers to meet and form an exciton. The exciton recombines
dissipating its energy through a photon.
Excitons formed from injected charges follow the statistics of forming three triplets
for every singlet. This limits the internal quantum efficiency of fluorescent emitters
30
to ∼25%[8] because the triplet states don’t emit. In 1998 Baldo et al. showed that
this limit could be raised to 100% by making use of phosphorescent emitters that
emit from triplet states[8, 9]. These highly efficient phosphorescent emitters are the
type that you’ll find in modern red and green subpixels of OLED displays2 . Almost
fifteen years later, Uoyama et al. showed that similar efficiencies can be realized in
fluorescent emitters that reverse-intersystem-cross triplets back to the singlet excited
state enabled by a small singlet-triplet gap ∆EST [119, 135]. This mechanism, termed
thermally activated delayed fluorescence (TADF), is shown schematically in figure 2-3
along with some of the molecules that exhibit it.
Figure 2-3: (a) Thermally activated delayed fluorescence (TADF) energy diagram.
(b) Some of the molecules that Uoyama et al. presented in their seminal TADF paper
that this figure is borrowed from (Reprinted with permission from [119]. Copyright
2012 Nature Publishing Group.).
To achieve such low ∆EST (≤ 100 meV), the single molecule is made up of both
donor and acceptor groups. While not necessarily faster or more efficient than phos2
Blue phosphorescent materials are, as of the time of writing, not yet efficient or stable
enough[100]
31
phorescent materials, TADF materials have a commercial advantage in that they
don’t make use of rare and expensive high spin–orbit coupling heavy metal cores like
the iridium found in Ir(ppy)3 3 .
3
The molecular structure of Ir(ppy)3 is shown in Appendix A
32
Chapter 3
The Sublimable Mask Lithographic
Process
he concept and process of sublimation mask lithography are described in
T
detail. While the idea and work here were developed independently, it is
important to mention that similar methods were investigated by IBM in
the late seventies[59] and early nineties[25] and by the Harvard Nanopore Group at
Harvard University more recently[65][14][48]. The work described here is the first time
that the dry nature of the method has been explored for use in patterning organic
semiconductor devices.
Each of the aforementioned works suggest or make use of a condensed gas as a
resist that is patterned, deposited on, and subsequently heated to induce lift-off. The
Harvard Nanopore Group used water as their resist material as it was inexpensive
and has a reasonably attainable freezing point. They demonstrated nanometer-scale
definition as shown in Figure 3-1. Unfortunately, water is one of the main materials
responsible for OLED degradation[16, 57, 110, 129] and an alternative sublimable
resist material is required.
33
Figure 3-1: An SEM image of The Harvard Nanopore Group’s demonstrated patterning of chromium using water ice resist. Reprinted with permission from [65].
Copyright 2005 American Chemical Society.
3.1
Theory of Practice
Below a certain pressure and above a certain temperature, materials sublime: their
thermodynamically distinct phase transitions directly from solid to vapor without
passing through a liquid phase. Deposition is the reverse transition. A phase diagram
showing this region is seen in figure 3-3. As previously mentioned, a dry process is
necessary for the production of organic semiconductors due to their sensitivity to
traditional solvents. Depositing a material on a substrate whose phase can be easily
changed between solid and vapor would make a versatile, clean and dry mask suitable
for patterning.
The heat delivery requirements Q to sublime a solid of mass m, with specific heat
capacity cV , and heat of sublimation hs are
Z
Tsub
Q=m
cV dT + mhS [38]
(3.1)
Ti
= m(cV ∆T + hS )
(3.2)
Because the change in temperature ∆T (the change from the initial temperature Ti to the sublimation point Tsub required to bring the mask to the sublimation
point is no more than 30◦ C, cV is much less than its room temperature value of 37.1
J·mol−1 K−1 [81] for the temperature regions of interest here and hS is 26.1 kJ·mol−1 [1]
for carbon dioxide, most of the heat going into the mask is required of sublimation
34
Vol. 24 • No. 46 • December 4 • 2012
D10488
www.advmat.de
Figure 3-2: Inside cover of December 2012 issue of Advanced Materials featuring
sublimation lithography[7]
rather than temperature increase, so
Q = m(cv ∆T + hS ) ≈ mhS
3.2
(3.3)
Process Flow
In practice, there are other necessary considerations. Heat exchange with gas molecules
at higher pressures causes the temperature to significantly fluctuate. The mask material would be very ineffective at the sublimation point due to the heat capacity
and heat of deposition. The system temperature must therefore be kept significantly
below this critical point for the resist material to be perform as intended. To raise
the temperature of the sublimation point, the mask is deposited at a higher pressure
to ensure that a higher portion of the gas is solidified. The pressure is later reduced
for the deposition of the thin film. Figure 3-4 shows an updated phase diagram to
reflect these details.
35
Pressure
Liquid
Solid
Gas
Temperature
Figure 3-3: General phase diagram. The arrow shows the ideal region of operation
with respect to sublimation lithography. Deposition of the mask takes place along
the low temperature region of the arrow and sublimation and lift-off take place along
the higher temperature region of the arrow.
With the substrate at higher pressures and low temperature (so as to be well
inside the solid portion of the phase diagram), the resist gas flows at and across
the substrate where much of it deposits on the surface (part (a) and (b) of figure
3-5. Next, the mask is selectively patterned exposing the substrate to allow for the
subsequently deposited neat films of the desired material to remain as intended post
lift-off (part (c)). The thin film is then grown over the mask at high-vacuum (part(d)).
Following this, the substrate temperature is brought up to sublime away the mask
which carries the undesired regions of thin film along with it (part(e)). These steps
can be performed again and again as needed to build devices that require multiple
patterned layers.
36
Pressure
Liquid
Solid
Deposition of
Mask
Deposition of
Thin-film
Sublimation/
Lift-off
Gas
Temperature
Figure 3-4: Phase diagram reflecting full process flow parameters. The defining of
the mask pattern is performed immediately prior to the deposition of the thin film.
(a)
Cooled
(d)
Glass
(b)
(c)
(e)
(f)
Organic 1
Phase-change resist
Organic 2
Metal
Figure 3-5: Simplified process flow for sublimation lithography. (not to scale) (a)
Begin with a cooled substrate to facilitate resist deposition. (b) Deposit resist. (c)
Selectively pattern resist. (d) Deposit desired thin film. (e) Lift-off resist leaving
patterned thin film. (f) Repeat as necessary to complete device.
37
38
Chapter 4
Sublimable Mask Patterning
he resist is patterned two ways in the sections that follow: local heating and
T
stamping. Photolithography is a third seemingly more elegant method and is
considered, but found to be impractical for the proof-of-concept material used.
4.1
Considerations for Demonstration
A small research-scale investigation is necessary to gauge the practicality of a fullscale sublimation patterning implementation. The basic principals of operation must
be demonstrated and fundamental limitations must be determined. If realizable, the
accuracy, repeatability, and resolution must be evaluated. Commercialization would
require evidence of 10 µm or smaller edge definition, as mentioned previously, as well
as some proof or arguments for yield and high throughput.
This work is a proof of principle for the fabrication of small-molecule organic
semiconductor devices. An existing thermal evaporator served as a testbed for the
original concept and all deviations therefrom. It was modified and refashioned as
necessary to accommodate improvements upon steps of the process or to approach
parts of the process from different angles. Full proof of commercializability would
require a full-scale evaporator built from the ground up with sublimable mask lift-off
patterning in mind, rather than a small research-scale system.
39
4.2
Experimental Setup
The experimental setup had to allow sufficient control of the process parameters
outlined in section 3.2. We have a thermal evaporator attached to a glovebox, designed
and built by Angstrom Engineering, that was easily modified to meet the process’
needs. The dry and inert nitrogen environment of the glovebox is crucial to avoid
water condensation on the substrate.
No substrate cooling was in place, so the rotating substrate holder was initially
replaced with a liquid nitrogen reservoir. A flange with tapped holes allowing for a
mounted substrate holder made up the bottom of the reservoir so the holder could
be in nearly direct contact with the liquid nitrogen. Temperatures as low as -165◦ C
were measured at pressures of 10−6 Torr with a type K thermocouple cemented to the
surface of a reference substrate. The difference between the surface temperature and
that of boiling liquid nitrogen (-196◦ C) is attributed to the thermal conductivities of
glass and steel. Indium foil inserted between each mating surface improves thermal
conductivity. More traditional cryogenic thermal greases were abandoned because of
their high photoluminescence that made it difficult to characterize patterned films
and the grease’s difficulty to cleanly work with in the glovebox.
Four feedthroughs (figure 4-1) allowed for observation and finer control of the
frozen mask: a feedthrough for the aforementioned thermocouple, a four-pin electrical feedthrough, a 1/4" steel tube feedthrough, and a universal serial bus (USB)
feedthrough. These allowed for temperature measurement, current sourcing (which
will be elaborated on in 4.2.1), gas flow for resist deposition, and use of USB peripherals, respectively. A USB hub with an attached light and webcam allowed for in situ
optical observation.
A substrate holder was machined out of aluminum. Holes in the corners allowed
for direct mounting to the bottom flange of the reservoir, while clips for securing
the substrate and any electrical probes were located along the perimeter, as seen in
figures 4-2 and 4-5. The center, though, was left in tact and smooth to minimize
thermal resistance.
40
Figure 4-1: The side of the evaporation chamber showing the feedthroughs and liquid
nitrogen reservoir.
Just below the holder, a copper tube is connected to the 1/4" steel tube feedthrough
and aimed at the substrate for CO2 gas flow. The camera is clipped to this tube out
of the way of organic material sources and thin film thickness sensors. A mass flow
controller is connected on the outside end of this tube for gas flow rate control.
Carbon Dioxide as a Resist
Carbon dioxide seemed like an obvious choice for a sublimation resist. It is chemically inert, and its sublimation point is easily accessible with a thermal evaporator
equipped with liquid nitrogen cooling and simple heaters. As phase data was not
readily available for the low pressure and temperature regime of this study, we extrapolated existing data with a Clapeyron equation-based thermodynamic fit to better
understand how CO2 would behave in our setup[112]. CO2 ’s phase diagram and our
extrapolation with pressure and temperature extremes of the experimental setup is
shown in figure 4-3.
41
Figure 4-2: Flange making up the bottom of the liquid nitrogen reservoir. The
substrate holding plate is bolted in with a glass sample tightly secured. The washermounted thermocouple can be seen towards the top of the image mounted with a
bolt.
The pressure of the chamber and temperature of the reservoir aren’t the sole
factors determining the phase of the CO2 resist. When a thin film is being deposited
onto the surface of the resist it must come to thermal equilibrium. Some of the resist
will warm up and sublime during this process. Based on the development in section
3.1, the thickness of resist required to survive the deposition of a particular thickness
of a thin film with enthalpy of fusion hf and vaporization hv can be approximated:
mCO2 (cv ∆T + hS )CO2 = mf ilm (cv ∆T + hv + hf )f ilm
(4.1)
Instead of using mass m, the density ρ can be used with the volume being made up
of area A and thickness t of the film to provide the thickness per unit area:
ρ=
m
→ m = ρAt
V
ρCO2 tCO2 (cv ∆T + hS )CO2 = ρf ilm tf ilm (cv ∆T + hv + hf )f ilm
42
(4.2)
(4.3)
Pressure / Torr
105
Liquid
CO2
100
H 2O
Solid
10-5
10-10
0
Solid
50
Vapor
Vapor
100 150 200 250 300 350
Temperature / K
Figure 4-3: Phase diagrams of CO2 and H2 O extrapolated (broken lines) from data
(solid lines) in references [83] and [120]. The “×” represents the process operating point of 77 K at 106 Torr. The curves are extrapolated using the Clapeyron
equation[112]. Figure from [2]
Then the thickness of CO2 given a desired thin film thickness is:
tCO2 =
ρf ilm tf ilm (cv ∆T + hv + hf )f ilm
ρCO2 (cv ∆T + hS )CO2
(4.4)
For 100 nm of silver, this is ∼3 µm and for 100 nm Alq3 , this is ∼400 nm (Alq3 ’s
hv , hf , and cv approximated with water’s values as thermal properties are not readily available). This approximation does not include the active cooling of the liquid
nitrogen reservoir or the heating from the evaporation source’s blackbody radiation.
The density of the resist depends on pressure, temperature[109], and gas flow rate
and, as mentioned in reference [46], a more amorphous resist avoids inhomogeneity
at the length scales of the crystalline domains and is preferred for greater resolution.
For the operating conditions in these experiments, the density is 1.51 ± 0.15 g cm−3 ;
43
see the Supporting Information of [2] for Allen Yin’s description of the interferometric
technique employed to measure density and film growth rates[32, 118].
Once the CO2 resist is patterned by selective sublimation, it is important to control
the partial pressure of CO2 in the chamber to prevent unwanted re-condensation of
CO2 vapor on patterned regions of the substrate. It is also possible to freeze other
impurity gases onto the substrate, notably H2 O, whose phase diagram is also shown
alongside that of CO2 in figure 4-3[120]. In previous studies of frozen CO2 films at
10−7 Torr, Gerakines et al. measured a water deposition rate of 2 nm h−1 [43]. At these
rates re-deposition must be considered in our experiments, but should ultimately be of
little consequence in high throughput manufacturing since the acceptable background
pressures of CO2 and H2 O increase with reduced takt time.
4.2.1
Patterning the Resist with Resistive Heating
We selectively patterned the CO2 mask by resistively heating a photolithographically
predefined conductor on the substrate surface. OLEDs always feature a bottom electrode layer such as the transparent conductor indium tin oxide (ITO), a thin metal,
or a transistor, so we considered ITO a realistic testbed.
The power P through any circuit element is given by
P = IV
(4.5)
where I is the current passing through the element and V is the voltage across that
element. By Ohm’s law,
V = IR
(4.6)
where R is the resistance of the circuit element. The resistance depends on the
resistivity ρ, the area A and the length l by the relation
R=ρ
l
A
(4.7)
From equation 3.3, we find the amount of heat needed to sublime away a target region
44
of CO2 of mass m is
Z
Q ≈ mhS =
t0
Z
t0
Z
2
I (t)Rdt =
I(t)V (t)dt =
0
0
0
t0
l
I 2 (t)ρ dt
A
(4.8)
As the ITO is essentially uniform in thickness D across the substrate and with predefined lines equal in width, this can be rewritten in terms of heat for a constant
thickness per unit width:
Q
mhS
D ≈D
= ρl
w
w
Z
t0
I 2 (t)dt
(4.9)
0
This leaves the current I(t) as the only active process parameter, which can be run
constant or pulsed accordingly to heat up and sublime away a desired mask region.
The ITO pattern used to test this selective sublimation method is displayed in
figure 4-4. Three 100 µm-wide lines allow for three 100 µm-per-side square subpixels
given a 100 µm-wide horizontally1 aligned top electrode. The six contact pads are
wider, both to facilitate electrical contact and so current passed through a line will
mostly heat the regions of highest resistance. The horizontal pad and center-right
island allow for contact with a cathode laid across the top of all three subpixels. In
this work, electrical contact in this work was made from the electrical feedthrough to
the substrate’s ITO pads in one of two ways: copper foil tape or small probe clips (as
seen in figure 4-5).
After a blanket deposition of CO2 at 100 Torr and ∼-165◦ C the chamber was
pumped down to ∼10−5 Torr. At this temperature, the resistance of a single 100
µm-wide line of the ITO used was about 500 Ω. 100 mA (current density of about
625 kA cm−2 ) was pulsed through the line with a Keithley 2400 sourcemeter while
the mask’s selective sublimation was observed via the camera. 20 nm of either tris(8hydroxyquinolinato)aluminium (Alq3 ), 4-(Dicyanomethylene)-2-tert-butyl-6-(1,1,7,7tetramethyljulolidin-4-yl-vinyl)-4H-pyran (DCJTB) (see Appendix A) or silver was
then thermally evaporated onto to the substrate (the choice of materials is explained
below). We then vented the chamber to atmospheric pressure, and heated the sample
1
Horizontal with respect to the orientation of figure 4-4
45
Figure 4-4: The pattern of the ITO used in resistive heating experiments. The black
regions are ITO; the white regions are bare glass.
with a thin ∼100 W kapton heater (Omega Engineering) sandwiched between the
substrate holder and liquid nitrogen reservoir. This forced the lift off of the CO2
mask and the undesired regions of the thin film. This was observed by video to
confirm complete sublimation. The pressure and temperature changes here reflect
those in figure 3-4.
Alq3 was chosen because it is a very well known OLED material (it was part
of Tang and VanSlyke’s first devices[117]). It is also has the advantage of obvious
photoluminescence. DCJTB is another highly photoluminescent material that has
been used as a dopant with Alq3 to produce red OLEDs[18]. An ultraviolet lamp
highlights the pattern resulting from the experimental procedure above, as in figure
4-7. Because the copper foil tape used to make electrical contact floats, it is poorly
cooled, so the adjacent regions show signs of CO2 mask degradation. Figure 4-8 shows
a micrograph of a similarly patterned line of silver. In both the images, the edges are
46
Figure 4-5: A mask of CO2 patterned for a 100 µm-wide line. The substrate holder
is rigged with probe-tipped clips for electrical contact.
fuzzy and the lines are not exactly 100 µm wide. In large part this is because of the
difficulties in fine calibration of power delivery by camera-mediated observation. A
specially designed system would ameliorate the problem and that is what led to an
alternative approach to defining the CO2 mask.
47
Copper
foil tape
CO2
Glass Substrate
160nm
ITO
Figure 4-6: In situ photograph and schematic representation of the resistive heating
method of patterning the frozen mask. Electrical contacts on the substrate allow
current to be driven at a current density of 625 kA cm2 through a 160 nm-thick strip
of ITO on the substrate. The red arrows in the schematic drawing indicate the flow
of current through the ITO enabled by affixed copper foil tape[2].
48
Figure 4-7: An exmaple of photoluminescence from a 100 µm-wide line of DCJTBdoped Alq3 (top), and photoluminescence(bottom-left) and a micrograph(bottomright) from a 100 µm-wide line of Alq3 patterned by resistive heating of ITO.
49
Figure 4-8: A micrograph from a 100 µm-wide line of silver patterned by resistive
heating of ITO
50
4.2.2
Patterning the Resist with a Stamp
We next investigated selective sublimation through direct stamping.
A micron-
featured stamp could heat specific regions of the mask through contact thereby patterning the mask. A stamp-specific version of the process flow is shown in figure 4-9.
We chose the epoxy-based SU-8 photoresists (inspired by [86]) to take advantage of
their versatility and rapid-prototyping capabilities: structures with thicknesses ranging from 1-600 µm with aspect ratios as high as 20:1[116] that can be made in a
matter of hours.
(a)
Cooled
(d)
(b)
(c)
(e)
(f)
(g)
Glass
Frozen CO 2
Organic Materials
Metal
SU-8 2150
Figure 4-9: Simplified process flow for sublimation lithography using a frozen CO2
resist (not to scale). (a) Cool substrate below 100 K. (b) Freeze on thin film of CO2 .
(c,d) Pattern CO2 film by heating selectively. Here a stamp is pressed against the
resist to remove some areas. (e) Deposit desired organic or metal thin film by thermal
evaporation. (f) Warm substrate to sublime CO2 , thus lifting off unwanted material.
(g) Repeat steps (a-f) as necessary to complete the device[2].
With stamping, the thermal energy required to sublime the CO2 mask is, from
equation 3.3,
Q = m(cV ∆Tstamp + hS ) ≈ mhS = mstamp cV stamp ∆Tstamp
(4.10)
Here ∆Tstamp does not reflect the difference of initial temperatures between the mask
and stamp, but rather the temperature change of the stamp after transferring energy
to sublime away the mask. The heat transfer is a function of the stamping pressure
and time, and if it is too great will sublime away necessary parts of the mask. To
51
develop this further, we could use Fourier’s law of conduction[70]
q = −k∇T
(4.11)
with the relation for thermal diffusivity[70]
α=
k
,
ρcV
(4.12)
where k is the thermal conductivity and ρ is the density. But until we can grow a
consistent resist layer and more accurately measure the thickness of the resist—and
thereby estimate the surface area—this exercise is not illuminative. An investigation
into the thickness of these resists under similar conditions can be found in Allen
Yin’s related work in the supplementary materials of the publication in Advanced
Materials[2].
A 2D simulation of the thermal diffusivity is shown in figure 4-10 and provides
a guide as to how the heat from the stamp moves once contact is made. The simulation expands on a unitless script written in MATLAB[99] to take into account
physical units, to match the diffusivity of the glass substrate as well as the length,
and time scales of interest. The method is a finite-difference approximation of the
heat diffusivity equation:
∂T
= α∇2 T
∂t
(4.13)
where α is as in equation 4.12, T is temperature, and t is time.
A few assumptions were made to reduce computation complexity and appropriate
boundary conditions were set. The bottom edge is set at liquid nitrogen’s boiling point
of 77 K, and all other edges have effectively pure vacuum insulation. The top middle
surface has a length held at 300 K to act as the stamp. The stamp temperature does
not drop and the thickness of the substrate is much less than it is in the experiments,
but should provide a sense of the movement of heat in the stamping progress. The
enthalpy of sublimation would also dissipate a great deal of heat from the stamp as
shown in equation 4.10. The iterations stop once the temperature no longer changes
52
Temperature / K
by more than 1% of the maximum temperature.
Figure 4-10: 2D simulation of thermal diffusivity in the stamping process.
We used SU-8 2150 obtained from Microchem to take advantage of its high thickness and aspect ratio potential—if the height of the features are not sufficiently
taller than the resist thickness, we’d risk subliming away the entire mask. The resist
was spun on solvent- and oxygen-plasma-cleaned silicon at 3000 rpm. Contact photolithography resolved ∼115 µm-tall pillars that tapered slightly after development in
propylene glycol methyl ether acetate (PGMEA). For these experiments, the tapering
is not so severe as to interfere with patterning as the resist thickness is on the order
of 50 µm. A micrograph of these features can be seen in figure 4-11. After growing a
blanket of CO2 on the cold substrate, the stamp is pressed into the frozen CO2 either
by hand or by a linear solenoid mounted on a rotational feedthrough. After stamping,
the thin film is deposited at pressures of ∼10−5 Torr. The mask is then lifted off,
ideally taking the undesired regions along with it. The results of these experiments
demonstrate the need for some optimization, as can be seen in figure 4-11.
Figure 4-11 suggests that either the heating in this configuration fails to fully
lift-off or—more likely— the stamping methods employed lack adequate feedback to
control the definition. Our hands may slip, and our solenoid-based mechanical system
53
Figure 4-11: A micrograph of the 115 µm-tall pillar SU-8 stamp(top) used to pattern
circles of Alq3 (bottom)
is not sufficient so as to prevent twisting and shifting that would thin or obfuscate the
mask. The particular stamp’s features may also have been too short to make contact
with the surface of the substrate without touching the rest of the mask.
Colder and with finer control
The next step was to gain better control of temperature, observation, and stamping.
This redesigned setup is shown schematically in figure 4-12. A cryogenic pump was
repurposed for use as a cooling source and all components are mounted onto it via an
54
oxygen-free high-conductivity (OFHC) copper rod. This gave us not only consistent
and reliable temperatures without having to refill boiling liquid nitrogen, but also
added temperature control in the previously unexplored region from around 10 K to
80 K. All cold parts are machined out of OFHC copper and indium foil is sandwiched
between all temperature-critical interfaces. A kapton encapsulated heater placed
in between the substrate and substrate holder provides adequate local heating for
encouraging lift-off without adding too much heat to the bulk thermal mass of the
apparatus. Below 80 K, a type K thermocouple is no longer of use so a silicon
thermal diode was attached to the copper substrate holder to approximately monitor
the temperature of the sample and a cryogenic temperature controller (Lakeshore
Cryotronics) was employed to manage operating temperature. To greatly improve
observation of our patterning as well as to investigate the yield of patterned thin films,
we added a digital microscope mounted on the stamp actuator. The microscope had
both white and 405 nm LEDs so photoluminescence imaging was attainable in situ
and we could also now record videos of lift-off in process.
CO2
Active
cooling
Cooled substrate
with underlying
heater
Thermal diode
Stamp &
microscope
high-vacuum
x & z stages
Evaporation source
Figure 4-12: Schematic of the updated experimental stamping setup showing the
process-critical components as described in the main text. The entire setup is located
within the standard thermal evaporation chamber (Angstrom Engineering). The
arrows by the stages indicate their respective direction and range of motion. Both the
stamp and the microscope used for positioning and stamping evaluation are located
on the smaller stage as indicated[2].
55
As seen in figure 4-12, the camera and stamp were mounted to two stages providing fine in situ control of motion. Stamping location and efficacy could be evaluated
in a matter of seconds and retried if insufficient stamping depth was observed. The
repurposed cryogenic pump’s compressor and cold head made for a lot of vibration
that made stamping during their operation impossible. Thus the compressor was
briefly turned off during the actual stamping and fine observations so that the vibrations did not interfere while the stamp and resist make contact or while recording
images. Because the cryopump’s cooling mechanism is closed-cycle rather than evaporative like the liquid nitrogen cooling it was replacing, longer cooling cycles reduced
throughput.
We also decided to redesign the stamp geometry in an attempt to demonstrate
features found in the current state of the art. Figure 4-13 shows this design meant to
meet the 325 pixels-per-inch in the tightest pitch commercially available product of the
time—the Apple iPhone Retina display[2]. The material and process for fabricating
the stamp remains the same.
Cooling down from room temperature takes a couple of hours depending on the
target temperature, but with effectively two cryogenic pumps in operation lower operating pressures as low as 10−7 Torr were attained. This isn’t necessarily good for
cleanliness because impurities, most importantly H2 O, could be sticking directly to
the substrate rather than just the main cryogenic pump as it can become just as cold
as the coldest object in the chamber. The steps that follow are analogous to those
from the earlier stamping discussion with the added complexity of controlling the
motorized linear stages. The X stage (150 mm travel) is moved to change between
stamping and observation mode whereas the Z stage (30 mm travel) controls stamping height or focus, respectively. Both of these components are controlled via USB
using proprietary software for each.
After flowing the CO2 onto the cooled substrate and carefully incrementing the
displacement of the stamp up to the resist the sublimable mask is formed, displayed
in part a of figure 4-14. As before, the organic film, Alq3 here, is sublimed as a
film onto the mask and through the mask onto the bare substrate. Running the
56
Figure 4-13: Micrograph of a region of the SU-8 stamp used to selectively sublime
regions of the CO2 resist for 78 µm-pitch patterns. In the center of the image is a
pillar that has been knocked over displaying the profile of the pillar[2].
kapton heater to overcome the cooling of the repurposed cryopump rapidly induces
sublimation of the resist layer. The subsequently deposited organic layer is lifted off
revealing the patterned film as in figure 4-14b. Figure 4-14c is a photoluminescence
micrograph of the obtained 325 pixel-per-inch density using a 380 nm LED with an
optical microscope (Carl Zeiss AG. Axioskop). For clarity, a longpass filter (Thorlabs,
Inc. FEL0450) was used to remove the pump from the image.
57
Figure 4-14: Optical micrograph (a) of a 78 µm pitch-patterned mask of CO2 defined
via contact with a stamp. The corresponding inset (b) shows the Alq3 thin film after
deposition and lift-off. Photoluminescence micrograph (c) from the same film[2]. A
longpass filter is used to lessen pump light’s presence in the image.
58
The edge definition and profile is of key importance ensuring that the processing
is suitable for device fabrication and that no roughness or debris is introduced by this
process. A false-color surface topography obtained using a Veeco Instruments Inc
optical profiler showing multiple pixels’ finer definition is displayed in figure 4-15(this
was also measured and confirmed on a Veeco-Tencor contact surface profilometer).
The bottom portion of the figure shows the cross section of a single pixel to examine
the smoothness of the remaining structured film. The observed topography shows no
problematic obstructions, roughness, or irregularities.
47nm
230µ
m
0nm
Height / nm
200µm
50
0
0
Distance / µm
50
Figure 4-15: A false-color topography obtained by an optical interferometer (top)
and a cross section of the same (bottom) detailing the profile of the patterned organic
pixels of figure 4-14[2].
It’s observed that the relative heat capacity of a stamp maintained at room temperature is more than sufficient to rapidly remove the frozen resist, taking mere
seconds to clear. To prevent abrasion and dust formation, the surface of the stamp
need not make contact with the hard substrate surface if a universal burn-off step is
performed to uniformly ‘etch’ the residual resist[48]. In this step, the temperature of
the sample and subsequently the resist is warmed up. A heater below the substrate
59
or one embedded in the substrate holder is used to combat the cooling mechanism.
This uniformly removes a suitable depth of all the resist such that none remains in
the areas where the desired thermally deposited film is to remain.
With low repeatability and yield from the experimental setup, it was not possible
to generate more than a few subpixels of successive layers at a time. Despite this,
the photoluminescence micrograph shown in figure 4-16 demonstrates the physical
capability of patterning more than one organic layer of different compounds one after
the other. The intended result was an array of Alq3 subpixels as shown in figure
4-14 followed by an array of Alq3 doped with the red dye DCJTB shifted 26 µm to
the side as would be seen in the first two color subpixel arrays of an RGB display.
Complete lift-off would not occur within the same area of the substrate for each color
film and we attributed this to insufficient flat levelling of the stamp with respect to
the substrate. The repeated red-green side-by-side subpixels seen only on the edge
of the substrate as seen in figure 4-16 support this idea because the edge would be
the only axis to make contact if the two interfaces do not share a normal. This is
not inherently a problem with using sublimable masks, but rather that the stamping
tools used in this study were not optimized. This is an essentially solved problem as
is seen by the success of soft lithography, nanoimprinting, and sophisticated contact
alignment[5, 15, 21, 45, 84, 105, 126].
60
Figure 4-16: Examples of red and green subpixels side-by-side. The left two images
show actual red and green subpixels next to each other whereas the right image
exemplifies some of the complications that arose in the attempts.
Free-floating thin film tensile strength
A repeatable yet quite unexpected phenomenon that was a common observation in
both successful and failed lift-off attempts was the ability of the organic thin film
to remain as a flexible continuous film. The organic film held together as it floated
freely folding away from the substrate with only a small portion of it still attached to
the substrate. We have recorded many instances of this in both white and UV light
micrographs and videos. Figure 4-17 shows a 50 nm-thick film of Alq3 lifting away
from the substrate yet remaining intact. It measures about 150 µm on a side. As
noted in chapter 2, it is a bit surprising at this scale that the weak Van der Waals
forces that bind organic molecules can hold the film together.
61
Figure 4-17: Photoluminescence micrograph of a free-floating 50 nm-thick film of Alq3
folding away from the substrate yet remaining intact. It measures about 150 µm on
a side.
4.3
OLEDs in a Cold Environment
To examine the impact of cold substrate temperatures on OLEDs, we built and tested
OLEDs on substrates cooled to 112 ± 24 K. We fabricated OLEDs using standard
shadow-masking techniques to demonstrate compatibility with the low temperature
substrates. Basic phosphorescent OLEDs employing fac tris(2-phenylpyridine) iridium (Ir(ppy)3 ) as the emitter were grown on ITO-coated glass substrates at similar
temperatures as the stamping process, but active areas were defined using traditional
shadow-masking techniques so as to evaluate the critical temperature-dependent parameter of using CO2 as a lift-off resist in producing organic optoelectronic devices.
The substrates were detergent, solvent, and plasma cleaned prior to device fabrication.
No CO2 was introduced in these experiments. N,N,N’,N’-tetrakis(4-methoxyphenyl)benzidine (MeO-TPD), 4,4’,4”-tris(carbazol-9-yl)triphenylamine (TCTA), 2,2’,2”-(1,3,5benzinetriyl)-tris(1-phenyl-1-H-benzimidazole) (TPBi), and lithium fluoride (LiF)/aluminum
(Al) were used as the hole transport, host, electron transport and cathode layers respectively.
The external quantum efficiency (EQE) versus current density, J, and electro62
luminescence spectrum of these devices are shown in figure 4-18. The best devices
on cold substrates yielded efficiencies comparable to the room temperature grown
control devices, suggesting that cooled substrate temperatures can be employed in
OLED fabrication without degradation of performance. The performance and yield
of the cold OLEDs was highly variable, however, and we observed a visible grey tint
in the hole transport layer due to a slight coarsening in the morphology. While morphological changes might occur during low temperature depositions, we attribute the
significant variation in device performance to the uncontrolled condensation of water
vapor or CO2 on our substrate surface during substrate cooling and the growth of
the thin films[57, 129]. This can be rectified by reducing the takt time from the ∼1 h
process used in our laboratory, and reducing the partial pressures of water and CO2
using cold traps[46].
In addition to addressing concerns of the organic layers’ growth under cold conditions, the transistor backplane of active-matrix displays must also withstand the low
temperatures of the process. To verify this, a small active-matrix OLED display was
removed from a digital photo frame, pumped down to high vacuum and cooled with
liquid nitrogen, and then replaced in its housing and reconnected to its driver. There
was no noticeable difference in pixel brightness, uniformity or operation aside from
the seal of the passivation glass coming loose—passivation being a manufacturing step
strictly after full device fabrication.
63
Room Temperature
12
10
EL Intensity / arb. units
External Quantum Efficiency / %
14
T = (112±24)K
8
6
4
2
(a)
0
1
0.8
100
10-1
Current Density / mA·cm-2
(b)
(c)
T = (112±24)K
0.6
100nm Al
0.8nm LiF
60nm TPBi
15nm TCTA:Ir(ppy)3
0.4
30nm MeO-TPD
ITO
0.2
0
101
Room Temperature
Glass
450 500 550 600 650 700 750
Wavelength / nm
Figure 4-18: External quantum efficiency versus current density of OLEDs grown
at T = 112 ± 24 K and room temperature (a). The normalized electroluminescence
spectrum is indistinguishable from the room temperature control device (b) and device
thin film stack (c) are also shown[2]
.
4.4
Patterning the Resist Photolithographically
While patterning the resist photolithographically would seem to be the most elegant, for CO2 this is impractical. A look at the absorption of solid CO2 [49] shows
strongest absorption at 2.7 µm and 4.3 µm. Overcoming both heat capacity and heat
of sublimation of actively cooled CO2 would require orders of magnitude more power
than is currently available at those wavelengths—the time it would take to pattern a
modern mother glass would be beyond unreasonable. Traditional photoresists have a
built-in chemical gain to cross-link and aren’t vaporized in their processing so require
much less incident energy. For research and development at a small scale, it could be
64
useful for rapid-prototyping purposes, but that has not yet been investigated due to
comparatively favorable cost and efficacy of the stamping method.
65
66
Chapter 5
Future Directions
he exploitation of a material’s sublimation point as a clearance mechanism
T
proved a suitable dry patterning method for organic thin films and metals.
However, reducing the energy requirement of the clearance dose would
enable more flexible pattern use, definition and possibly reduce takt time. Below is
a discussion of a couple of additional organic thin film patterning methods.
5.1
Combustion Lithography
An extensive amount of work has been done in exploring the use of nitrocellulose as
a “self-developing resist” in the 1980s an early 1990s[30, 40, 41, 53, 69, 87]. It was
used to pattern metals[42], stated to be compatible with dry etches, and was touted
as being a dry process. If this is true, it may serve as a patterning method for organic
semiconductors. It would share the dry nature of a sublimable mask, but have an
advantage of a much lower clearance dose; there is already built-up energy that makes
up the explosive nature of the material. An example of patterned nitrocellulose from
reference [41] is shown in figure 5-1.
There is a question, though, of combustion by-product residues leftover. These
could interfere with thin film devices easily causing shorts, pinholes, or other defects.
Geis et al. considered two possible degradation pathways, once of which resulted in
solid compounds as opposed to simple gases[41]. The other consideration is whether
67
Figure 5-1: SEM micrograph of nitrocellulose on silicon patterned via an argon ion
beam by Geis et al. in 1983. Reprinted with permission from [41]. Copyright 1983
American Vacuum Society.
multiple layers can be patterned without destroying underlying layers from the patterning and lift-off combustion.
5.2
Magnetic Resists
The trouble with grand physical or chemical changes in a resist is that they often
come along with unwanted by-products or side effects: with traditional photolithography the substrate can be left with unevaporated solvents or insufficiently rinsed
resist; with sublimation lithography one has to watch out for unwanted water or
particulates sticking to the substrate surface under cooled conditions; and with combustion lithography, residual char or water can reduce yield. How great would it be
to selectively remove your resist material without these concerns? Enter magnetic
materials.
This suggested approach adapts and expands upon past work using magnetic
materials[12, 58] as well as some concepts from xerography[17, 58, 121]. It focuses
explicitly on solvent-free processing for organic LEDs.
Like sublimable resists, the concept is simple with a basic understanding of lithography and the physical phenomenon exploited; in this case, magnetism. The process
68
flow is schematically displayed in figure 5-2. The challenging portion of the process
is arranging the resist where you want it. Resist could be positioned via a scanning
solenoid head like that on a hard disk drive, but that would make for a very slow
serial process. Influenced by xerography and demonstrated by Jacobs et al.[58], the
material could be put down in pattern electrostatically, then post thin film deposition, removed via magnetic field. Jacobs et al. used a charging stamp to deposit
electrolets on a polymer that positioned printer toner, iron beads, and iron oxide as
shown in figure 5-3.
Figure 5-2: Suggested process flow for using magnetic materials as dry lift-off resists.
(not to scale) (a) Begin with clean substrate (b) Apply magnetic resist where subsequent thin film is not desired (c) Deposit thin film (d) Remove resist and unwanted
thin film areas with a magnet or other strong magnetic field (e) Repeat as necessary
to complete all layers of desired device.
Another option would be to sit the substrate on a magnetic mold, of sorts. This
would be a surface with an intricate pattern of micron-scale field lines arranging a
blanket-deposited magnetic powder. This would be akin to placing the powder on
a thin sheet of glass mounted on a flexible vinyl magnet. These magnets are made
with alternating magnetic domains as shown in figure 5-4 thus giving you alternating
fields. For OLED applications, a much more advanced version of this field pattern
would be required and there would be trade-offs and limitations based on field shape,
strength, and substrate thickness.
69
Figure 5-3: Jacobs et al. stamped elecrolets on PMMA to position the above materials. Reprinted with permission from [58]. Copyright 2001 American Association for
the Advancement of Science.
Figure 5-4: Magnet domain arrangement of a flexible vinyl magnet and the resulting
field lines. Image from [123]
70
Chapter 6
Conclusion
O
LEDs must be capable of smaller features on a larger standard mother glass
with higher yields if they are to be cost-competitive with other display
technologies. Current masking techniques have fallen short of meeting
those goals, but sublimable masks are capable of filling the gap. We have used CO2
to show that a sublimable mask can pattern thin films of both organic semiconductors
and metals. We demonstrated two methods of mask definition: resistive heating of
the underlying electrode and stamping. Both processes avoid the mask damage that
plagues the FMM technique, but, unlike inkjet and other wet-processing methods,
don’t constrain the deposition materials.
With the concept proved, the next step is to show that sublimable masks can
deliver at a large scale. Commercial production requires accurate alignment to the
substrate (each of the three subpixels is made of different materials for the higher
efficiency direct-emission subpixel scheme), high device yield, and low takt time. For
sumblimable masks, the resistive heating method would require a restructuring of
the driving backplane of the display, but 1-3 µm alignment accuracy has already
been shown for gravure cylinder and roll printing of electronics[66][89], and a cylinder
etched deeply enough to stamp a CO2 mask could work like the stamp in section
4.2.2. Cooling and heating could be pipelined to preserve takt time, but without
test devices fabricated, yield cannot at this point be evaluated. Still, there is no
reason to believe that substrate cooling during device growth would negatively affect
71
OLEDs, which have been known to operate at cryogenic temperatures[26], and we
foresee neither fundamental limitation to scaling this process up to larger substrates,
nor to achieving smaller feature sizes.
There are other applications for this process and variants of it besides thin film
definition for OLEDs. For one, in processes involving the removal of film or structures
intended to be free-floating, they could be fabricated on top of a blanket of frozen
CO2 that could later be sublimed similar to the observations of section 4.2.2. The
same could be done for materials grown or synthesized, by chemical vapor deposition
or similar, that might otherwise adhere strongly to the walls of a chamber. A spatial
inverse of the cooling and mask process involving only briefly cooling the lift-off
mask region of the substrate for a temporary mask that sublimes as the temperature
reequilibrates is another approach. This is similar to what Jacobs et al. did as shown
in figure 5-3 by sticking a masking material down to specific areas with charge[58],
instead exploiting thermal differences.
While pixel density as high as 325 pixels-per-inch has been demonstrated in this
work, there is no reason to believe this is a fundamental limit—especially knowing
that nanometer-scale patterning has been demonstrated by others with an electron
beam paired with frozen water resist[37, 46, 47, 48, 65]. Patterning organics at the
nanoscale may be possible pairing phase-change resists with nanoimprinting techniques. Employing a micro-featured stamp roller pipelined with the necessary cooling apparatuses, phase-change resist patterning should allow for scaling of parallel
patterning beyond what the current technologies offer.
72
Bibliography
[1] D. Ambrose, R. M. Stephenson, and S. Malanowski. Handbook of the thermodynamics of organic compounds. 1987.
[2] Apple - iPhone 4 - learn about the high-resolution retina display.
http://www.apple.com/iphone/features/retina-display.html, 2010.
[3] Apple - iPhone 5s - technical specifications. https://www.apple.com/iphone5s/specs/, 2014.
[4] Apple — ipad — compare ipad models. http://www.apple.com/ipad/compare/,
2014.
[5] Michael D Austin, Haixiong Ge, Wei Wu, Mingtao Li, Zhaoning Yu, D Wasserman, SA Lyon, and Stephen Y Chou. Fabrication of 5 nm linewidth and
14 nm pitch features by nanoimprint lithography. Applied Physics Letters,
84(26):5299–5301, 2004.
[6] Matthias E Bahlke, Hiroshi A Mendoza, Daniel T Ashall, Allen S Yin, and
Marc A Baldo. Dry lithography of large-area, thin-film organic semiconductors
using frozen co2 resists. Advanced Materials, 24(46):6136–6140, 2012.
[7] Matthias E Bahlke, Hiroshi A Mendoza, Daniel T Ashall, Allen S Yin, and
Marc A Baldo. Organic semiconductors: Dry lithography of large-area, thin-film
organic semiconductors using frozen co2 resists (adv. mater. 46/2012). Advanced
Materials, 24(46):6116–6116, 2012.
[8] M. A. Baldo, D. F. O’Brien, Y. You, A. Shoustikov, S. Sibley, M. E. Thompson, and S. R. Forrest. Highly efficient phosphorescent emission from organic
electroluminescent devices. Nature, 395(6698):151–154, 1998.
[9] M. A. Baldo, M. E. Thompson, and S. R. Forrest. High-efficiency fluorescent organic light-emitting devices using a phosphorescent sensitizer. Nature,
403(6771):750–753, February 2000.
[10] MA Baldo, VG Kozlov, PE Burrows, SR Forrest, VS Ban, B Koene, and
ME Thompson. Low pressure organic vapor phase deposition of small molecular weight organic light emitting device structures. Applied physics letters,
71(21):3033–3035, 1997.
73
[11] MA Baldo, S Lamansky, PE Burrows, ME Thompson, and SR Forrest. Very
high-efficiency green organic light-emitting devices based on electrophosphorescence. Applied Physics Letters, 75(1):4–6, 1999.
[12] Amos Bardea and Ron Naaman. Magnetolithography: From bottom-up route
to high throughput. Small, 5(3):316319, 2009.
[13] Moon Bo-kyung.
Samsung display order
eration
amoled
line
equipment
anytime
http://english.etnews.com/device/2781223 1304.html.
the a3 6th gensoon
|
etnews.
[14] Daniel Branton, Jene A. Golovchenko, Gavin M. King, Warren J. MoberlyChan, and Gregor M. Schrmann. Lift-off patterning processing employing
energetically-stimulated local ..., April 2009. Filing Date: Dec 9, 2004 U.S.
Classification: 216/41.
[15] Timothy Burgin, Vi-En Choong, and George Maracas. Large area submicrometer contact printing using a contact aligner. Langmuir, 16(12):5371–5375, 2000.
[16] PE Burrows, V Bulovic, SR Forrest, LS Sapochak, DM McCarty, and
ME Thompson. Reliability and degradation of organic light emitting devices.
Applied Physics Letters, 65(23):2922–2924, 1994.
[17] Chester F Carlson. Electrophotography, October 6 1942. US Patent 2,297,691.
[18] C. H. Chen, C. W. Tang, J. Shi, and K. P. Klubek. Improved red dopants for
organic electroluminescent devices. In Die Makromolekulare Chemie. Macromolecular symposia, page 4958, 1997.
[19] Jianglong Chen, Valrie Leblanc, Sung Hoon Kang, Paul J Benning, David Schut,
Marc A Baldo, Martin A Schmidt, and Vladimir Bulovic. High definition digital fabrication of active organic devices by molecular jet printing. Advanced
Functional Materials, 17(15):2722, August 2007.
[20] Roger Cheng. Samsung mobile display to boost 2011 production. wsj.com,
September 2010.
[21] Stephen Y Chou, Peter R Krauss, and Preston J Renstrom. Nanoimprint lithography. Journal of Vacuum Science & Technology B, 14(6):4129–4133, 1996.
[22] Kyuha Chung, Namdeog Kim, Joonhoo Choi, Changwoong Chu, and Jongmoo Huh. 70.1: Invited paper: Large-sized full color amoled tv: Advancements
and issues. In SID Symposium Digest of Technical Papers, volume 37, pages
1958–1963. Wiley Online Library, 2006.
[23] Jennifer Colegrove. DisplaySearch quarterly OLED shipment and forecast report. Technical report, DisplaySearch, 2010.
74
[24] Daniel N Congreve, Jiye Lee, Nicholas J Thompson, Eric Hontz, Shane R Yost,
Philip D Reusswig, Matthias E Bahlke, Sebastian Reineke, Troy Van Voorhis,
and Marc A Baldo. External quantum efficiency above 100% in a singletexciton-fission–based organic photovoltaic cell. Science, 340(6130):334–337,
2013.
[25] JJ Cuomo, CR Guarnieri, KL Saenger, and DS Yee. Selective deposition with
”Dry” vaporizable Lift-Off mask. IBM Technical Disclosure Bulletin, 35(1A),
June 1992.
[26] Brian W. D’Andrade, James Esler, and Julie J. Brown. Organic light-emitting
device operational stability at cryogenic temperatures. Synthetic Metals, 156(56):405–408, March 2006.
[27] Robert D. Deegan. Pattern formation in drying drops. Physical Review E,
61(1):475, January 2000.
[28] John A. DeFranco, Bradley S. Schmidt, Michal Lipson, and George G.
Malliaras. Photolithographic patterning of organic electronic materials. Organic Electronics, 7(1):22–28, February 2006.
[29] Caleb Denison.
Lg 77-inch ultra hd oled tv — digital trends.
http://www.digitaltrends.com/home-theater/lg-77-inch-ultra-hd-oled-tvgallery/.
[30] T. F. Deutsch and M. W. Geis. Selfdeveloping UV photoresist using excimer
laser exposure. Journal of Applied Physics, 54(12):7201–7204, December 1983.
[31] David Hsieh in DisplaySearch, Flat Panel Displays on April 2, and
2013 | 0 Comments. 2013 AMOLED revenues to pass $10b | DisplaySearch blog. http://www.displaysearchblog.com/2013/04/2013-amoledrevenues-to-pass-10b/.
[32] M Domingo, C Millán, M Aı̄ Satorre, and J Cantó. Thin film thickness measurement by double laser interferometry. In Optical Metrology, pages 66164A–
66164A. International Society for Optics and Photonics, 2007.
[33] OLED MICRODISPLAYS | eMagin.
microdisplays/.
http://www.emagin.com/oled-
[34] Ayataka Endo, Mai Ogasawara, Atsushi Takahashi, Daisuke Yokoyama,
Yoshimine Kato, and Chihaya Adachi. Thermally activated delayed fluorescence from Sn4+Porphyrin complexes and their application to organic light
emitting diodes a novel mechanism for electroluminescence. Advanced Materials, 21(47):48024806, 2009.
[35] V Fock. Z. physik 61, 126 (1930); jc slater. Phys. Rev, 35:210, 1930.
75
[36] Charlotte Froese Fischer. General hartree-fock program. Computer physics
communications, 43(3):355–365, 1987.
[37] Jules A Gardener and Jene A Golovchenko. Ice-assisted electron beam lithography of graphene. Nanotechnology, 23(18):185302, 2012.
[38] D. R Gaskell. Introduction to the Thermodynamics of Materials. Hemisphere
Pub, 2003.
[39] Bernard Geffroy, Philippe le Roy, and Christophe Prat. Organic light-emitting
diode (OLED) technology: materials, devices and display technologies. Polymer
International, 55(6):572–582, 2006.
[40] M. W. Geis, J. N. Randall, T. F. Deutsch, P. D. DeGraff, K. E. Krohn, and
L. A. Stern. Selfdeveloping resist with submicrometer resolution and processing
stability. Applied Physics Letters, 43(1):74–76, July 1983.
[41] M. W. Geis, J. N. Randall, T. F. Deutsch, N. N. Efremow, J. P. Donnelly, and
J. D. Woodhouse. Nitrocellulose as a self-developing resist with submicrometer
resolution and processing stability. Journal of Vacuum Science & Technology
B: Microelectronics and Nanometer Structures, 1(4):1178–1181, 1983.
[42] M. W. Geis, J. N. Randall, R. W. Mountain, J. D. Woodhouse, E. I. Bromley,
D. K. Astolfi, and N. P. Economou. Nitrocellulose as a positive or negative selfdeveloping resist. Journal of Vacuum Science & Technology B: Microelectronics
and Nanometer Structures, 3(1):343–346, 1985.
[43] PA Gerakines, WA Schutte, JM Greenberg, and Ewine F van Dishoeck. The
infrared band strengths of h2o, co and co2 in laboratory simulations of astrophysical ice mixtures. arXiv preprint astro-ph/9409076, 1994.
[44] Patrick Görrn, Michelle Sander, Jens Meyer, Michael Kröger, Eike Becker, H-H
Johannes, Wolfgang Kowalsky, and Thomas Riedl. Towards see-through displays: fully transparent thin-film transistors driving transparent organic lightemitting diodes. Advanced Materials, 18(6):738–741, 2006.
[45] L Jay Guo. Nanoimprint lithography: methods and material requirements.
Advanced Materials, 19(4):495–513, 2007.
[46] Anpan Han, John Chervinsky, Daniel Branton, and Jene A Golovchenko. An ice
lithography instrument. Review of Scientific Instruments, 82(6):065110–065110,
2011.
[47] Anpan Han, Aaron Kuan, Jene Golovchenko, and Daniel Branton. Nanopatterning on nonplanar and fragile substrates with ice resists. Nano letters,
12(2):1018–1021, 2012.
76
[48] Anpan Han, Dimitar Vlassarev, Jenny Wang, Jene A. Golovchenko, and Daniel
Branton. Ice lithography for nanodevices. Nano Letters, 10(12):5056–5059,
December 2010.
[49] Gary B Hansen. The infrared absorption spectrum of carbon dioxide ice from
1.8 to 333 m. Journal of Geophysical Research, 102(E9):21,569–21,587, 1997.
[50] DR Hartree and W Hartree. Results of calculations of atomic wave functions.
iii. results for be, ca and hg. Proceedings of the Royal Society of London. Series
A-Mathematical and Physical Sciences, 149(867):210–231, 1935.
[51] Hiroshi Hayase. DisplaySearch quarterly OLED shipment and forecast report.
Technical report, DisplaySearch, 2010.
[52] TR Hebner, CC Wu, D Marcy, MH Lu, and JC Sturm. Ink-jet printing of
doped polymers for organic light emitting devices. Applied Physics Letters,
72(5):519–521, 1998.
[53] D. Henderson, J. C. White, H. G. Craighead, and I. Adesida. Selfdeveloping photoresist using a vacuum ultraviolet f2 excimer laser exposure. Applied
Physics Letters, 46(9):900–902, May 1985.
[54] Pierre Hohenberg and Walter Kohn. Inhomogeneous electron gas. Physical
review, 136(3B):B864, 1964.
[55] Htc one (m8) specs and reviews — htc
http://www.htc.com/www/smartphones/htc-one-m8/.
additional
support.
[56] Ha Soo Hwang, Alexander A. Zakhidov, Jin-Kyun Lee, Xavier Andr, John A.
DeFranco, Hon Hang Fong, Andrew B. Holmes, George G. Malliaras, and
Christopher K. Ober. Dry photolithographic patterning process for organic
electronic devices using supercritical carbon dioxide as a solvent. Journal of
Materials Chemistry, 18(26):3087, 2008.
[57] Takeshi Ikeda, Hideyuki Murata, Yoshiki Kinoshita, Junichi Shike, Yoshikazu
Ikeda, and Masahiro Kitano. Enhanced stability of organic light-emitting devices fabricated under ultra-high vacuum condition. Chemical physics letters,
426(1):111–114, 2006.
[58] Heiko O Jacobs and George M Whitesides. Submicrometer patterning of charge
in thin-film electrets. Science, 291(5509):1763–1766, 2001.
[59] WL Johnson, RB Laibowitz, and CC Tsuei. Condensed gas, in situ lithography.
IBM Technical Disclosure Bulletin, 20(9), February 1978.
[60] Zakya H Kafafi. Organic electroluminescence. CRC Press, 2005.
[61] DM Karnakis, T Lippert, N Ichinose, S Kawanishi, and H Fukumura. Laser
induced molecular transfer using ablation of a triazeno-polymer. Applied surface
science, 127:781–786, 1998.
77
[62] Kateeva introduces YIELDjet an inkjet printing manufacturing equipment solution for mass producing flexible and large-size OLEDs | kateeva. http://kateeva.com/press-full/kateeva-introduces-yieldjetan-inkjet-printing-manufacturing-equipment...
[63] Hye Dong Kim, Jae Kyeong Jeong, Hyun-Joong Chung, and Yeon-Gon Mo.
22.1: Invited paper: Technological challenges for large-size amoled display. In
SID Symposium Digest of Technical Papers, volume 39, pages 291–294. Wiley
Online Library, 2008.
[64] Miyoung Kim. Samsung unit sees firm smartphone market. Reuters, May 2010.
[65] Gavin M. King, Gregor Schrmann, Daniel Branton, and Jene A. Golovchenko.
Nanometer patterning with ice. Nano Letters, 5(6):1157–1160, June 2005.
[66] E. Kitazume, K. Takeshita, K. Murata, Y. Qian, Y. Abe, M. Yokoo, K. Oota,
and T. Taguchi. 41.2: Development of polymer Light-Emitting diode (PLED)
displays using the relief printing method. SID Symposium Digest of Technical
Papers, 37(1):1467–1470, June 2006.
[67] M Knupfer. Exciton binding energies in organic semiconductors.
Physics A, 77(5):623–626, 2003.
Applied
[68] Walter Kohn and Lu Jeu Sham. Self-consistent equations including exchange
and correlation effects. Physical Review, 140(4A):A1133, 1965.
[69] C. E. Kosmidis and C. D. Skordoulis. Dopant-enhanced ablation of nitrocellulose by a nitrogen laser. Applied Physics A, 56(1):64–68, January 1993.
[70] S. Kou. Transport phenomena and materials processing. Wiley-Interscience,
1996.
[71] SL Lai, MY Chan, MK Fung, CS Lee, LS Hung, and ST Lee. Applications of
ytterbium in organic light-emitting devices as high performance and transparent
electrodes. Chemical physics letters, 366(1):128–133, 2002.
[72] Jin-Kyun Lee, Margarita Chatzichristidi, Alexander A. Zakhidov, Priscilla G.
Taylor, John A. DeFranco, Ha Soo Hwang, Hon Hang Fong, Andrew B. Holmes,
George G. Malliaras, and Christopher K. Ober. Acid-sensitive semiperfluoroalkyl resorcinarene: An imaging material for organic electronics. Journal of
the American Chemical Society, 130(35):11564–11565, September 2008.
[73] Jun Yeob Lee and SEONG TAEK Lee. Laser-induced thermal imaging of polymer light-emitting materials on poly (3, 4-ethylenedioxythiophene): Silane holetransport layer. Advanced Materials, 16(1):51–54, 2004.
[74] Jung-Ah Lee. In the lead–and trying to stay there. wsj.com, November 2010.
78
[75] Jung-Ah Lee. Samsung group to invest $38.3 billion in 2011. wsj.com, January
2011.
[76] Kwan Hee Lee, Seoung-Yoon Ryu, Jang Hyuk Kwon, Sang Wook Kim, and
Ho Kyoon Chung. 9.3: 2.2 qcif full color transparent amoled display. In SID
Symposium Digest of Technical Papers, volume 34, pages 104–107. Wiley Online
Library, 2003.
[77] Seong Taek Lee, Byung Doo Chin, Mu Hyun Kim, Tae Min Kang, Myung Won
Song, Jae Ho Lee, Hye Dong Kim, Ho Kyoon Chung, Martin B Wolk, Erika
Bellmann, et al. 29.3: A novel patterning method for full-color organic lightemitting devices: Laser induced thermal imaging (liti). In SID Symposium
Digest of Technical Papers, volume 35, pages 1008–1011. Wiley Online Library,
2004.
[78] Seong Taek Lee, Jun Yeob Lee, Mu Hyun Kim, Min Chul Suh, Tae Min Kang,
Yun Jin Choi, Joon Young Park, Jang Hyuk Kwon, Ho Kyoon Chung, John
Baetzold, et al. 21.3: A new patterning method for full-color polymer lightemitting devices: Laser induced thermal imaging (liti). In SID Symposium
Digest of Technical Papers, volume 33, pages 784–787. Wiley Online Library,
2002.
[79] Gilbert N Lewis and M Kasha. Phosphorescence and the triplet state. Journal
of the American Chemical Society, 66(12):2100–2116, 1944.
[80] 55-inch OLED TV panel — lgd newsroom.
products-solutions/tv/652.
http://lgdnewsroom.com/
[81] D. R. Lide. CRC handbook of chemistry and physics. (Internet version 2011).
Boca Raton, FL: CRC, 2011.
[82] Jiin-Jou Lih, Ching-Ian Chao, and Chung-Chun Lee. Novel pixel design for
high-resolution amoled displays with a shadow mask. Journal of the Society for
Information Display, 15(1):3–7, 2007.
[83] J. P Martin Trusler. Equation of state for solid phase i of carbon dioxide valid
for temperatures up to 800 k and pressures up to 12 GPa. Journal of Physical
and Chemical Reference Data, 40(4):043105–043105–19, December 2011.
[84] Bruno Michel, André Bernard, Alexander Bietsch, Emmanuel Delamarche,
Matthias Geissler, David Juncker, Hannes Kind, J-P Renault, H Rothuizen,
Heinz Schmid, et al. Printing meets lithography: soft approaches to highresolution patterning. IBM Journal of Research and Development, 45(5):697–
719, 2001.
[85] Robert S Mulliken. Electronic structures of polyatomic molecules and valence.
ii. general considerations. Physical Review, 41(1):49, 1932.
79
[86] Apoorva Murarka, Corinne Packard, Frank Yaul, Jeffrey Lang, and Vladimir
Bulovic. Micro-contact printed mems. In Micro Electro Mechanical Systems
(MEMS), 2011 IEEE 24th International Conference on, pages 292–295. IEEE,
2011.
[87] R. Mhle and G. Gtz. Ion beam exposure characteristics of nitrocellulose. Nuclear
Instruments and Methods in Physics Research Section B: Beam Interactions
with Materials and Atoms, 46(14):347–349, February 1990.
[88] Hajime Nakanotani, Kensuke Masui, Junichi Nishide, Takumi Shibata, and
Chihaya Adachi. Promising operational stability of high-efficiency organic lightemitting diodes based on thermally activated delayed fluorescence. Scientific
Reports, 3, July 2013.
[89] Seung-Hee Nam, Soon-Sung Yoo, Youn-Gyoung Chang, Nam-Kook Kim, YunHoo Kook, Jinook Kim, Chang-Dong Kim, Inbyeong Kang, and In-Jae Chung.
43.4: Enhancement of roll printing accuracy for TFT-LCD. SID Symposium
Digest of Technical Papers, 39(1):648–650, May 2008.
[90] Tetsuro Ohtsuka, Masahide Tsukamoto, and Mitsuharu Tsuchiya. Liquid crystal matrix display. Japanese Journal of Applied Physics, 12(No. 3):371–378,
1973.
[91] OLED displays to hit sales of $44b by 2019, with growth in TVs and mobile devices. http://venturebeat.com/2012/08/15/oled-displays-to-hit-44bby-2019-with-growth-in-tvs-and-mobile-devices/.
[92] William O’Mara, Robert B Herring, and Lee P Hunt. Handbook of semiconductor silicon technology. Crest Publishing House, 2007.
[93] Panasonic
develops
56-inch
4k2k
oled
panel.
http://panasonic.co.jp/corp/news/official.data/data.dir/2013/01/en1301093/en130109-3.html.
[94] Robert G Parr. Density functional theory. Annual Review of Physical Chemistry, 34(1):631–656, 1983.
[95] G Parthasarathy, Chihaya Adachi, PE Burrows, and Stephen R Forrest. Highefficiency transparent organic light-emitting devices. Applied Physics Letters,
76(15):2128–2130, 2000.
[96] Dora. Pesach and Abraham. Marmur. Marangoni effects in the spreading of
liquid mixtures on a solid. Langmuir, 3(4):519–524, July 1987.
[97] M Pfeiffer, SR Forrest, X Zhou, and K Leo. A low drive voltage, transparent,
metal-free n–i–p electrophosphorescent light emitting diode. Organic electronics, 4(1):21–26, 2003.
80
[98] Martin Pope and Charles E Swenberg. Electronic processes in organic crystals,
volume 1. Clarendon Press Oxford, 1982.
[99] Sathyanarayan Rao.
Finite difference method to solve heat diffusion equation in two dimensions - matlab file exchange.
http:
//www.mathworks.com/matlabcentral/fileexchange/42604-finitedifference-method-to-solve-heat-diffusion-equation-in-twodimensions/content/heat_2d.m, 2013.
[100] Sebastian Reineke. Organic light-emitting diodes: Phosphorescence meets its
match. Nature Photonics, 8(4):269–270, 2014.
[101] Sebastian Reineke, Frank Lindner, Gregor Schwartz, Nico Seidler, Karsten
Walzer, Björn Lüssem, and Karl Leo. White organic light-emitting diodes with
fluorescent tube efficiency. Nature, 459(7244):234–238, 2009.
[102] Sebastian Reineke, Gregor Schwartz, Karsten Walzer, Meiken Falke, and Karl
Leo. Highly phosphorescent organic mixed films: The effect of aggregation on
triplet-triplet annihilation. Applied Physics Letters, 94(16):163305, 2009.
[103] Sebastian Reineke, Gregor Schwartz, Karsten Walzer, and Karl Leo. Reduced
efficiency roll-off in phosphorescent organic light emitting diodes by suppression
of triplet-triplet annihilation. Applied Physics Letters, 91(12):123508, 2007.
[104] Sebastian Reineke, Karsten Walzer, and Karl Leo. Triplet-exciton quenching in
organic phosphorescent light-emitting diodes with ir-based emitters. Physical
Review B, 75(12):125328, 2007.
[105] John A Rogers and Ralph G Nuzzo. Recent progress in soft lithography. Materials today, 8(2):50–56, 2005.
[106] Samsung galaxy s5. http://www.samsung.com/global/microsite/galaxys5/specs.html,
2014.
[107] Samsung galaxy and apple iPhone driving 55% increase in mobile phone
display revenues in 2013, according to NPD DisplaySearch - DisplaySearch.
http://www.displaysearch.com/cps/rde/xchg/displaysearch/hs.xsl/
130822_samsung_galaxy_and_apple_iphone_driving_5_increase_in_
mobile_phone_display_revenues.asp.
[108] Seongkeum amoled tv era smd-lgd, 55” amoled tv panels developed | etnews.com. http://www.etnews.com/201109160141, September 2011.
[109] MÁ Satorre, M Domingo, C Millán, R Luna, R Vilaplana, and C Santonja.
Density of ch4, n2 and co2 ices at different temperatures of deposition. Planetary
and Space Science, 56(13):1748–1752, 2008.
81
[110] Michel Schaer, Frank Nüesch, Detlef Berner, William Leo, and Libero Zuppiroli. Water vapor and oxygen degradation mechanisms in organic light emitting diodes. Advanced Functional Materials, 11(2):116–121, 2001.
[111] Francis Weston Sears. Principles of physics, volume 1. Addison-Wesley Press,
Incorporated, 1948.
[112] Francis Weston Sears. An introduction to thermodynamics: the kinetic theory
of gases, and statistical mechanics. Addison-Wesley Pub. Co., 1953.
[113] James R. Sheats. Manufacturing and commercialization issues in organic electronics. Journal of Materials Research, 19(7):1974–1989, 2004.
[114] Max Shtein, Peter Peumans, Jay B. Benziger, and Stephen R. Forrest. Micropatterning of small molecular weight organic semiconductor thin films using
organic vapor phase deposition. Journal of Applied Physics, 93(7):4005, 2003.
[115] John C Slater. Atomic shielding constants. Physical Review, 36(1):57, 1930.
[116] SU-8 resists, SU-8 2000, & SU-8 3000 resists - MicroChem. http://www.
microchem.com/products/su_eight.htm.
[117] C. W. Tang and S. A. VanSlyke. Organic electroluminescent diodes. Applied
Physics Letters, 51(12):913–915, 1987.
[118] KE Tempelmeyer and DW Mills Jr. Refractive index of carbon dioxide cryodeposit. Journal of Applied Physics, 39(6):2968–2969, 2003.
[119] Hiroki Uoyama, Kenichi Goushi, Katsuyuki Shizu, Hiroko Nomura, and Chihaya Adachi. Highly efficient organic light-emitting diodes from delayed fluorescence. Nature, 492(7428):234–238, December 2012.
[120] Wolfgang Wagner, A. Saul, and A. Pruss. International equations for the pressure along the melting and along the sublimation curve of ordinary water substance. Journal of Physical and Chemical Reference Data, 23(3):515–527, May
1994.
[121] Edgar M Williams. The physics and technology of xerographic processes. Krieger
Publishing Company, 1993.
[122] Dustin Winters, Joel D. Shore, and Steven A. Van Slyke. White-emitting organic electroluminescent device with color filters and reflective layer for causing
colored light constructive interference, May 2004. U.S. Classification 313/504,
313/505, 313/113; International Classification H01L51/52, H05B33/28,
H05B33/12, C09K11/06, H01L51/30, H01L51/50, H01L27/32, H05B33/26,
H01L51/00;
Cooperative Classification C09K2211/18,
H01L51/005,
C09K2211/1003, H01L51/0059, C09K2211/1018, H01L51/0077, Y02B20/181,
C09K11/06, H01L51/0084, H01L27/322, C09K2211/10, H01L51/0052,
H01L51/5262; European Classification H01L51/52D, C09K11/06.
82
[123] Refrigerator Magnets | University of Wisconsin-Madison MRSEC Nanostructured Interfaces.
http://education.mrsec.wisc.edu/background/
fridgemag/.
[124] Shuanghong Wu, Masaki Aonuma, Qisheng Zhang, Shuping Huang, Tetsuya
Nakagawa, Kazuhiro Kuwabara, and Chihaya Adachi. High-efficiency deepblue organic light-emitting diodes based on a thermally activated delayed fluorescence emitter. Journal of Materials Chemistry C, 2(3):421–424, December
2013.
[125] xdrawchem chemistry editor, January 2011.
[126] Younan Xia and George M Whitesides. Soft lithography. Annual review of
materials science, 28(1):153–184, 1998.
[127] Hong Xiao. Introduction to semiconductor manufacturing technology, volume 16. Prentice Hall Upper Saddle River, NJ, 2001.
[128] Asuka Yamamori, Sachiko Hayashi, Toshiki Koyama, and Yoshio Taniguchi.
Transparent organic light-emitting diodes using metal acethylacetonate complexes as an electron injective buffer layer. Applied Physics Letters, 78(21):3343–
3345, 2001.
[129] H Yamamoto, J Brooks, MS Weaver, JJ Brown, T Murakami, and H Murata.
Improved initial drop in operational lifetime of blue phosphorescent organic
light emitting device fabricated under ultra high vacuum condition. Applied
Physics Letters, 99(3):033301, 2011.
[130] Kiyoshi Yoneda and Kenichi Shibata. Color display apparatus having electroluminescence elements, May 2002. U.S. Classification 313/506, 313/112;
International Classification H05B33/28, H01L27/32, H01J1/74, H05B33/14,
H01J1/62, H01L51/50, G09F9/30, H05B33/12, H05B33/26, H01L51/52;
Cooperative Classification H01L27/322, H01L51/5284, H01L2251/5315,
H01L27/3244, H01L51/5203, H01J1/74, H01L51/5206; European Classification
H01J1/74, H01L27/32M2, H01L27/32C6, H01L51/52B2.
[131] Kyoung Soo Yook, Soon Ok Jeon, Chul Woong Joo, and Jun Yeob Lee. Color
stable and interlayer free hybrid white organic light-emitting diodes using an
area divided pixel structure. Synthetic Metals, 159(17-18):1778–1781, September 2009.
[132] Sae Youn Lee, Takuma Yasuda, Hiroko Nomura, and Chihaya Adachi. Highefficiency organic light-emitting diodes utilizing thermally activated delayed fluorescence from triazine-based donoracceptor hybrid molecules. Applied Physics
Letters, 101(9):093306–093306–4, August 2012.
[133] Alexander A. Zakhidov, Jin-Kyun Lee, John A. DeFranco, Hon Hang Fong,
Priscilla G. Taylor, Margarita Chatzichristidi, Christopher K. Ober, and
83
George G. Malliaras. Orthogonal processing: A new strategy for organic electronics. Chemical Science, 2(6):1178, 2011.
[134] Alexander A. Zakhidov, Jin-Kyun Lee, Hon Hang Fong, John A. DeFranco,
Margarita Chatzichristidi, Priscilla G. Taylor, Christopher K. Ober, and
George G. Malliaras. Hydrofluoroethers as orthogonal solvents for the chemical
processing of organic electronic materials. Advanced Materials, 20(18):3481–
3484, 2008.
[135] Qisheng Zhang, Bo Li, Shuping Huang, Hiroko Nomura, Hiroyuki Tanaka, and
Chihaya Adachi. Efficient blue organic light-emitting diodes employing thermally activated delayed fluorescence. Nature Photonics, advance online publication, March 2014.
[136] X Zhou, M Pfeiffer, JS Huang, J Blochwitz-Nimoth, DS Qin, A Werner, J Drechsel, B Maennig, and K Leo. Low-voltage inverted transparent vacuum deposited
organic light-emitting diodes using electrical doping. Applied physics letters,
81(5):922–924, 2002.
84
Appendix A
Selected Molecules Used in this
Work
N
O
O
N
Al
N
O
Figure A-1: Aluminium tris(quinolin-8-olate) (Alq3 )[125]
O=C=O
Figure A-2: Carbon dioxide (CO2 )
85
N
O
N
N
Figure A-3: 4-(Dicyanomethylene)-2-tert-butyl-6-(1,1,7,7-tetramethyljulolidin-4-ylvinyl)-4H-pyran (DCJTB)[125]
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
Figure A-4: SU-8 Photoresist[125]
86
O
Figure A-5: Tris[2-phenylpyridinato-C2 ,N ]iridium(III)(Ir(ppy)3 )[11]
87
88
Appendix B
Triple Points of Some Gases
Substance
Temperature (K)
Pressure (Torr)
Acetylene
192.4
900.08
Ammonia
195.40
45.57
Argon
83.81
516.79
Butane
134.6
52.50
Graphite
4765
75996.46
Carbon dioxide
216.55
3877.83
Carbon monoxide
68.10
115.28
Chloroform
175.43
6.53
Deuterium
18.63
128.26
Ethane
89.89
60.01
Ethanol
150
32.25
Ethylene
104.0
0.90
Formic acid
281.40
16.50
lambda point
2.19
38.25
Hexafluoroethane
173.08
199.52
Hydrogen
13.84
52.80
Hydrogen chloride
158.96
104.26
Iodine
386.65
90.53
89
Substance
Temperature (K)
Pressure (Torr)
Isobutane
113.55
14.61
Mercury
234.2
12.38
Methane
90.68
87.76
Neon
24.57
324.03
Nitric oxide
109.50
164.41
Nitrogen
63.18
94.51
Nitrous oxide
182.34
658.93
Oxygen
54.36
1.14
Palladium
1825
26.25
Platinum
2045
15.00
Sulfur dioxide
197.69
12.53
Titanium
1941
39.75
Uranium hexafluoride
337.17
1137.85
Water
273.16
4.59
Xenon
161.3
611.30
Zinc
692.65
0.49
90
Appendix C
MATLAB Code
Below is some of the code I used for data analysis during this work. Some lines may
be commented and modified to handle intricacies of unique data sets. Additionally,
functionality may have been added without proper documentation. That being said,
it should still be pretty readable on its own.
constants.m
1
% Load fundamental constants
2
3
% h = 4.135667516 E -15; % eV * s
4
h = 6.62606957 E -34; % J * s
5
c = 2.99792458 E17 ;
6
% c = 2.99792458 E8 ; % in m / s
7
q = 1.60217646 E -19;
% in Coulombs
8
kb = 8.617332478 E -5;
% in ev / K
% in nm / s
OLEDEQE.m
1
function [ OLEDdata , extEff , extEffcurrent ] = OLEDEQE ( IVdatafile , ELspectrumFile , pdser ,
idrive , mask , figurenumber , reducedark , normalize , plot ) ;
2
%
3
% OLED EQE calculation based on the following article :
4
%
5
% Forrest , S . , Bradley , D . and Thompson , M . (2003) , Measuring the
6
% Efficiency of Organic Light - Emitting Devices . Advanced Materials ,
7
% 15: 1043
1 0 4 8 . doi : 10.1002/ adma .200302151
91
8
%
9
% Author : Matthias E Bahlke
10
%
11
% Example usage : >> OLEDEQE ( ’ S1D2C . TXT ’ , ’ spectrum . txt ’ , ’4574 ’ ,0 , ’ mixed ’ ,9 ,1 ,0 ,1) ;
12
%
13
% IVdatafile is the file containing the IV of the OLED and photocurrent
14
% from the diode
15
% ELspec trumFil e is the file containing the EL spectrum
16
% pdser is the serial number of the photodetector
17
% idrive is a binary / boolean value to select whether it is VI or IV sweep
18
% ( yes for IV ) . Binary boolean value
19
% mask is whether the small or mix device cathode mask was used . Valid
20
% inputs : ’ small ’ ’ mixed ’
21
% figurenumber is the number of the figure intended to create the plot
22
% reducedark subtracts the DC photocurrent from the photodiode from all
23
% values . Binary boolean value
24
% normalize normalizes the EQE at the current density value specified
25
% below . Binary boolean value
26
% plot , whether or not to plot . Binary boolean value
27
28
29
switch mask
case ’ mixed ’
30
devicenumber = regexp ( IVdatafile , ’\ SD (\ d ) \ S ’ , ’ tokens ’) ;
31
32
devicenumber = str2num ( cell2mat ( devicenumber {1}) ) ;
33
if mod ( devicenumber ,2) > 0
activearea = (.1/2) ^2* pi ;
34
else
35
activearea = (.144/2) ^2* pi ;
36
end
37
case ’ dresden ’
38
activearea = (0.25475) ^2;
39
case ’ small ’
40
activearea = (0.05) ^2* pi ;
41
case ’ newsmall ’
42
activearea = (.144/2) ^2* pi ;
43
case ’ huge ’
44
activearea = (0.5) ^2* pi ;
45
case ’ samsung ’
46
activearea = (.2) ^2;
47
case ’ squares ’
48
activearea = (.1) ^2;
49
otherwise
50
activearea = mask ;
51
52
end
92
53
54
55
56
57
58
59
% Set geometrical factor f , old value of 0.26
60
f = 0.21;
61
62
% Build filename for specified PD serial number and subsequently load
63
% responsivity data
64
65
if isunix
pdrespfile = strcat ( ’/ home / matthias / Dropbox / matlab / PDresp / ’ , pdser , ’. txt ’) ;
66
67
end
68
69
if ismac
pdrespfile = strcat ( ’/ Users / matthias / Dropbox / matlab / PDresp / ’ , pdser , ’. txt ’) ;
70
71
end
72
73
74
75
pdresp = load ( pdrespfile , ’ ’) ;
76
77
% Load necessary fundamental constants ; BE SURE TO CHECK UNITS !!!
78
constants ;
79
80
% Load OLED EL Spectral data
81
% global ELspectrum ;
82
% ELspectrum = importdata ( ELspectrumFile , ’ ’) ;
83
84
% Application of the calibration file should remove wavelengths
85
% incompatible with the photodetector respon sivities
86
87
ELspectrum = c o rr ec tS p ec tr um ( E Lspectru mFile ) ; % , ’/ home / matthias / Dropbox / matlab /
o ou sb co r re ct io n . txt ’)
88
89
% Load OLED and photodetector sweep data
90
global OLEDdata ;
91
OLEDdata = importdata ( IVdatafile ) ; % ’
92
OLEDdata = OLEDdata . data ;
’ ,5) ;%22) ;% , headersize ) ;%) ;
93
94
% Add numbering if necessary
95
96
if 0
93
97
dataindex = (1:1: length ( OLEDdata (: ,1) ) ) ’;
98
OLEDdata = [ dataindex OLEDdata ]
99
end
100
101
102
103
% Reduce dark current of PD
104
105
if reducedark
OLEDdata (: ,4) = OLEDdata (: ,4) - OLEDdata (1 ,4) ;
106
107
end
108
109
110
% Swap coloumns of OLEDdata if it is current driven
111
if idrive
112
currenthold = OLEDdata (: ,2) ;
113
voltagehold = OLEDdata (: ,3) ;
114
OLEDdata (: ,2) = voltagehold ;
115
OLEDdata (: ,3) = currenthold ;
116
end
117
118
119
120
% Interpolate pd responsivity data to enable operation with other spectral
121
% data
122
global pdrespint ;
123
pdrespint = interp1 ( pdresp (: ,1) , pdresp (: ,2) , ELspectrum (: ,1) ) ;
124
125
% Perform integration accounting for EL and PDresp spectral overlap
126
global RfactorInt ;
127
RfactorInt = trapz ( ELspectrum (: ,2) .* pdrespint ) / trapz ( ELspectrum (: ,1) .* ELspectrum (: ,2)
);
128
129
% Using center wavelength , LESS ACCURATE !
130
% RfactorInt = (1/624) * trapz ( ELspectrum (: ,2) .* pdrespint ) / trapz ( ELspectrum (: ,2) ) ;
131
132
% Calculate the external quantum efficiency
133
global extEff ;
134
extEff = 100* q * abs ( OLEDdata (: ,4) ) ./( h * c * f * OLEDdata (: ,3) * RfactorInt ) ;
135
136
extEffcurrent = OLEDdata (: ,3) *1000/ activearea ;
137
138
% Supress " negative data ignored " messages on command window
139
warning off MATLAB : Axes : N e g a t i v e D a t a I n L o g A x i s
140
94
141
142
143
if normalize > 0
144
% find ’ normalize ’ array position
145
[ min_difference , array _positi on ] = min ( abs ( OLEDdata (: ,3) *1000/ activearea normalize ) ) ;
146
% divide extEff by value at above array position
147
extEff = extEff /( extEff ( array _positio n ) ) ;
148
149
end
150
151
152
% Plot the EQE of the device on a semilogx with respect to current density
153
154
if plot
155
figure ( figurenumber ) ;
156
[ upperPath , deepestFolder , ~] = fileparts ( pwd ) ;
157
DeviceName = strcat ( IVdatafile ,{ ’ - ’} , deepestFolder ) ;
158
DeviceName = ’ hi ’;
159
semilogx ( OLEDdata (: ,3) *1000/ activearea , extEff , ’ LineWidth ’ ,2 , ’ DisplayName ’ ,
DeviceName ) ;
xlim ([2 E -2 ,100]) ;
160
161
xlabel ( ’ Current [ mA cm ^{ -2}] ’) ;
162
if normalize
163
ylim ([0 ,1.2]) ;
164
ylabel ( ’ Normalized External Quantum Efficiency ’) ;
else
165
ylim ([0 ,16]) ;
166
ylabel ( ’ External Quantum Efficiency [%] ’) ;
167
168
end
169
title ( ’ External Quantum Efficiency vs . Current ’) ;
170
end
plotOLEDEQEs.m
1
function [] = plotOLEDEQEs ( ELspectrumFile , pdser , firstLetter , idrive , mask , figurenum
, reducedark , normalize , write2file )
2
3
% ELspec trumFile is the file containing the e l e c t r o l u m i n e s c e n c e from the
4
%
5
% pdser is the serial number of the photodetector
6
% firstLetter is the prefix of the set of files considered for analysis
7
% idrive is set to ’1 ’ if the driving current is the first column of the
8
%
9
% mask is set to ’ mixed ’ if the cathode is both large and small areas ,
photodetector
data file , ’0 ’ otherwise
95
10
%
anything else otherwise ( ’ small ’ for example )
11
% figurenum is the number of the figure to plot to
12
% reducedark is set to ’1 ’ if you ’ d like to substract the dark current from
13
%
14
% normalize is set to the current density to normalize the EQE to , ’0 ’ otherwise
15
% write2file is set to ’1 ’ to output the calculated EQE to file
16
%
17
%
all photodetector currents considered
18
19
20
21
% for example for data of the form : S1D1_0V . txt ,
22
% plotOLEDEQEs ( ’ S2D1 / spec00007 . txt ’ , ’4448 ’ , ’ S ’ ,0 , ’ small ’ ,63 ,1 ,0 ,0) ;
23
24
25
26
fileFormat = strcat ( firstLetter , ’ *. TXT ’) ;
27
28
files = dir ( fileFormat ) ;
29
numberoffiles = size ( files ) ;
30
if numberoffiles (1) == 0
fileFormat = strcat ( firstLetter , ’ *. txt ’) ;
31
files = dir ( fileFormat ) ;
32
33
end
34
numberoffiles = size ( files ) ;
35
if numberoffiles (1) == 0
36
fileFormat = strcat ( firstLetter , ’ *. lvm ’) ;
37
files = dir ( fileFormat ) ;
38
end
39
numberoffiles = size ( files ) ;
40
if numberoffiles (1) == 0
’ The format doesn ’ ’t seem to match ’
41
42
end
43
44
45
figure ( figurenum ) ;
46
47
hold on ;
48
for i =1: length ( files )
49
50
% eval ([ ’ load ’ files ( i ) . name ’ - ascii ’]) ;
51
52
IVdatafile = files ( i ) . name ;
53
96
% [ OLEDdata , extEff ] = OLEDEQE ( IVdatafile , ELspectrumFile , pdser , idrive , mask ,
54
figurenum , reducedark , normalize , plot
55
[ OLEDdata , extEff , extEffcurrent ] = OLEDEQE ( IVdatafile , ELspectrumFile , pdser ,
56
idrive , mask , figurenum , reducedark , normalize ,0) ;
57
if write2file
58
eqeFileName = strcat ( IVdatafile , ’. eqe ’) ;
59
60
fileColumns = horzcat ( extEffcurrent , extEff ) ;
61
dlmwrite ( eqeFileName , fileColumns , ’\ t ’) ;
end
62
63
64
semilogx ( extEffcurrent , extEff , ’ LineWidth ’ ,2 , ’ DisplayName ’ , IVdatafile , ’ tag ’ , ’p ’) ;
65
% plot ( wv , loadFactor * EQE , ’ LineWidth ’ , 2 , ’ DisplayName ’ , EQEmeas , ’ tag ’ , ’p ’) ;
66
67
end
68
69
xlim ([2 E -2 ,100]) ;
70
xlabel ( ’ Current [ mA cm ^{ -2}] ’) ;
71
if normalize > 0
72
ylim ([0 ,1.2]) ;
73
ylabel ( ’ Normalized External Quantum Efficiency ’) ;
74
else
ylim ([0 ,16]) ;
75
ylabel ( ’ External Quantum Efficiency [%] ’) ;
76
77
end
78
79
[ upperPath , deepestFolder , ~] = fileparts ( pwd ) ;
80
title ( strcat ( ’ External Quantum Efficiency vs . Current ’ ,{ ’ - ’} , deepestFolder ) ) ;
81
82
% Set a different color for each line
83
84
lh = findobj ( gcf , ’ tag ’ , ’p ’) ;
85
cm = jet ( length ( files ) ) ;
86
cs = size ( cm ,1) ;
87
for i =1: length ( lh )
set ( lh ( i ) , ’ color ’ , cm ( rem (i -1 , cs ) +1 ,:) ) ;
88
89
end
90
91
set ( gca , ’ XScale ’ , ’ log ’)
correctSpectrum.m
1
function [ c o r r e c t e d S p e c t r u m ] = co rr e ct Sp ec t ru m ( rawData , c orrectio nFile ) ;
2
97
3
if nargin < 2
4
% correc tionFile = ’/ home / matthias / Dropbox / matlab / o ou sb co r re ct i on . txt ’;
5
corre ctionFi le = ’ oo u sb co rr e ct io n . txt ’;
6
7
end
8
9
10
% Load data files
11
u n c o r r e c t e d S p e c t r u m = load ( rawData ) ;
12
c o r r e c t i o n S p e c t r u m = load ( correc tionFile ) ;
13
14
15
% Interpolate the file to be corrected to the correction file
16
17
u n c o r r e c t e d S p e c t r u m I n t = interp1 ( u n c o r r e c t e d S p e c t r u m (: ,1) , u n c o r r e c t e d S p e c t r u m (: ,2) ,
c o r r e c t i o n S p e c t r u m (: ,1) ) ;
18
19
% Apply the calibration
20
21
22
s pe ct ra l Pr od uc t = c o r r e c t i o n S p e c t r u m (: ,2) .* u n c o r r e c t e d S p e c t r u m I n t ;
23
24
c o r r e c t e d S p e c t r u m = c o r r e c t i o n S p e c t r u m (: ,2) .* u n c o r r e c t e d S p e c t r u m I n t ;
25
26
c o r r e c t e d S p e c t r u m = horzcat ( c o r r e c t i o n S p e c t r u m (: ,1) , sp ec tr a lP ro du c t ) ;
plotivs.m
1
function [] = plotivs ( firstLetter , mask , headerlines , figurenum , idrive )
2
3
%
4
%
5
%
6
%
7
%
8
%
9
%
10
%
11
%
12
%
Plot IVs of a directory
plotivs ( firstLetter , mask , headerlines , figurenum , idrive )
13
14
15
16
17
fileFormat = strcat ( firstLetter , ’ *. TXT ’) ;
98
18
19
files = dir ( fileFormat ) ;
20
numberoffiles = size ( files ) ;
21
if numberoffiles (1) == 0
22
fileFormat = strcat ( firstLetter , ’ *.* ’) ;
23
files = dir ( fileFormat ) ;
24
end
25
26
27
figure ( figurenum ) ;
28
29
hold on ;
30
for i =1: length ( files )
31
% eval ([ ’ load ’ files ( i ) . name ’ - ascii ’]) ;
32
33
sweepname = files ( i ) . name ;
34
35
36
37
% Parse mask size and set active area
38
39
switch mask
case ’ mixed ’
40
devicenumber = regexp ( IVdatafile , ’\ SD (\ d ) \ S ’ , ’ tokens ’) ;
41
42
devicenumber = str2num ( cell2mat ( devicenumber {1}) ) ;
43
if mod ( devicenumber ,2) > 0
activearea = (.1/2) ^2* pi ;
44
else
45
activearea = (.144/2) ^2* pi ;
46
end
47
case ’ dresden ’
48
activearea = (0.25475) ^2;
49
case ’ small ’
50
activearea = (0.05) ^2* pi ;
51
case ’ newsmall ’
52
activearea = (.144/2) ^2* pi ;
53
case ’ huge ’
54
activearea = (0.5) ^2* pi ;
55
case ’ samsung ’
56
activearea = (.2) ^2;
57
case ’ squares ’
58
activearea = (.1) ^2;
59
otherwise
60
activearea = mask ;
61
62
end
99
63
ivdata = importdata ( files ( i ) . name ) ; % , ’\t ’ , headerlines ) ;
64
65
66
% Swap coloumns of OLEDdata if it is current driven
67
if idrive
currenthold = ivdata . data (: ,2) ;
68
69
voltagehold = ivdata . data (: ,3) ;
70
ivdata . data (: ,2) = voltagehold ;
ivdata . data (: ,3) = currenthold ;
71
end
72
73
semilogy ( ivdata . data (: ,2) , abs ( ivdata . data (: ,3) *1000/ activearea ) , ’ LineWidth ’ , 2 , ’
74
DisplayName ’ , sweepname , ’ tag ’ , ’p ’) ;
75
76
end
77
78
79
80
% Set a different color for each line
81
82
lh = findobj ( gcf , ’ tag ’ , ’p ’) ;
83
cm = jet ( length ( files ) ) ;
84
cs = size ( cm ,1) ;
85
for i =1: length ( lh )
set ( lh ( i ) , ’ color ’ , cm ( rem (i -1 , cs ) +1 ,:) ) ;
86
87
end
88
89
90
xlabel ( ’ Voltage [ V ] ’) ;
91
ylabel ( ’ Current [ mA cm ^{ -2}] ’) ;
92
% ylim ([0 ,150]) ;
93
[ upperPath , deepestFolder , ~] = fileparts ( pwd ) ;
94
set ( gca , ’ YScale ’ , ’ log ’)
95
96
title ( strcat ( ’{ JVs from } ’ , deepestFolder ) ) ;
97
hold off ;
98
99
100
101
return ;
loadTransient.m
1
function [] = loadTransient ( filename , figurenum , expnum , s u b t r a c t B a s e l i ne )
2
%
100
3
% Load transient data file and calculate lifetime
4
%
5
% Author : Matthias E Bahlke
6
%
7
%
8
%
Script will plot once for user to determine each area of interest
9
%
Usage :
10
%
for filename ’ TEK000 . csv ’
trans01 = loadTransient ( ’ TEK000 . csv ’ ,5 ,1) ;
11
%
12
%
13
14
15
figure ( figurenum ) ;
16
hold on ;
17
18
transientData = load ( filename ) ;
19
20
21
22
23
% Determine baseline by averaging where t < 0
24
25
tbaseline = find ( transientData (: ,1) ==0) ;
26
baseline = mean ( transientData (1: tbaseline -100 ,2) ) ;
27
28
29
if s u b t r a c t B a s e l i n e
transientData (: ,2) = transientData (: ,2) - baseline ;
30
31
else % add baseline
transientData (: ,2) = transientData (: ,2) + baseline ;
32
33
end
34
35
36
% plot to determine range of interest
37
warning ( ’ off ’ , ’ MATLAB : Axes : N e g a t i v e D a t a I n L o g A x i s ’) ;
38
39
plot ( transientData (: ,1) , abs ( transientData (: ,2) ) ) ; % - baseline ) ;
40
41
xlabel ( ’ Time / s ’) ;
42
ylabel ( ’ Counts / arb . units ’) ;
43
% ylim ([ -.3 ,1.2]) ;
44
xlim ([0 , max ( transientData (: ,1) ) ]) ;
45
set ( gca , ’ YScale ’ , ’ log ’)
46
% set ( gcf , ’ renderer ’ , ’ zbuffer ’) ;
47
101
48
colors = [ ’r ’ , ’k ’ , ’g ’ , ’b ’ ];
49
50
for i =1: expnum
51
% ask user for range of interest
52
53
54
%
tstart = input ( ’ When to start ? ’) ;
55
%
tfinish = input ( ’ When to stop ? ’) ;
56
title ( ’ Select start and stop of range of interest ’) ;
57
[x , y ] = ginput (2) ;
58
tstart = x (1) ;
59
tfinish = x (2) ;
60
61
% To match nearby time choices , reduce to 2 significant digits
62
% tstart = round ( tstart /(10^( floor ( log10 ( tstart ) ) -1) ) ) *(10^( floor ( log10 ( tstart ) )
63
% tfinish = round ( tfinish /(10^( floor ( log10 ( tfinish ) ) -1) ) ) *(10^( floor ( log10 ( tfinish
-1) ) ;
) ) -1) ) ;
64
65
% find
closest value to those clicked
66
67
for k =1: size ( transientData (: ,1) )
if tstart - transientData (k ,1) <= 0
68
istart = k ;
69
break ;
70
end
71
72
end
73
74
for k =1: size ( transientData (: ,1) )
if tfinish - transientData (k ,1) <= 0
75
ifinish = k ;
76
break ;
77
end
78
79
end
80
81
82
83
% do fit and grab time constant
84
85
decayfit = fit ( transientData ( istart : ifinish ,1) , transientData ( istart : ifinish ,2) , ’
exp1 ’) ; % - baseline , ’ exp1 ’) ;
86
% figure (41) ;
87
decayplot = plot ( decayfit ) ;
88
fitvalues = coeffvalues ( decayfit ) ;
89
timeconst = -1/ fitvalues (2) ;
102
set ( decayplot , ’ LineWidth ’ ,1 , ’ color ’ , colors ( i ) , ’ DisplayName ’ , strcat ( ’ fitted curve
90
’ , int2str ( i ) ) ) ;
91
title ( ’ Place time constant label ’) ;
92
gtext ( strcat ( ’\ tau_ { ’ , int2str ( i ) , ’} = ’ , En gineers Style ( timeconst ) , ’s ’) ) ; %
93
set ( gca , ’ YScale ’ , ’ log ’) ;
num2str ( timeconst ) ,’ s ’) ) ;
94
95
end
96
97
% set ( gcf , ’ renderer ’ , ’ painters ’) ;
98
99
100
% Add some figure labels and limits
101
102
xlabel ( ’ Time / s ’) ;
103
ylabel ( ’ Counts / arb . units ’) ;
104
xlim ([0 , max ( transientData (: ,1) ) ]) ;
105
[ upperPath , deepestFolder , ~] = fileparts ( pwd ) ;
106
title ( strcat ( ’ Transient from ’ ,{ ’ - ’} , deepestFolder ) ) ;
EngineersStyle.m
1
function Str = Engine ersStyl e ( x )
2
3
% %%%
4
%
5
% Edited from Jan Simon and Walter Roberson
6
% http :// www . mathworks . com / matlabcentral / answers /892
7
%
8
%
9
%
10
11
Exponent = 3 * floor ( log10 ( x ) / 3) ;
12
y = x / (10 ^ Exponent ) ;
13
ExpValue = [9 , 6 , 3 , 0 , -3 , -6 , -9];
14
ExpName = { ’G ’ , ’M ’ , ’k ’ , ’ ’ , ’m ’ , ’\ mu ’ , ’n ’ };
15
ExpIndex = ( Exponent == ExpValue ) ;
16
if any ( ExpIndex )
17
18
% Found in the list :
Str = sprintf ( ’ %.2 f % s ’ , y , ExpName { ExpIndex }) ;
else
% Fallback : Show the numeric value
19
% EDITED : Walter refined ’%d ’ to ’%+04 d ’:
20
Str = sprintf ( ’% fe %+04 d ’ , y , Exponent ) ;
21
end
103
plqy.m
1
function [] = plqy ( blankfile , samplefile , material , figurenum )
2
3
%
4
%
5
%
6
%
7
%
8
%
9
%
10
%
plqy ( ’ NPB_blank . txt ’ , ’ NPB_sample . txt ’ , ’ PS : NPB ’ ,45) ;
11
12
blank = c o rr ec tS p ec tr um ( blankfile ) ;
13
sample = c or re ct S pe ct r um ( samplefile ) ;
14
15
figure ( figurenum ) ;
16
hold on ;
17
18
plot ( blank (: ,1) , blank (: ,2) , ’ DisplayName ’ , ’ Blank ’) ;
19
plot ( sample (: ,1) , sample (: ,2) , ’ DisplayName ’ , ’ Sample ’) ;
20
set ( gca , ’ YScale ’ , ’ log ’) ;
21
22
title ( ’ Select pump start and stop of range of interest ’) ;
23
[x , y ] = ginput (2) ;
24
tstart = x (1) ;
25
tfinish = x (2) ;
26
27
for k =1: size ( blank (: ,1) )
if tstart - blank (k ,1) <= 0
28
29
pstart = k ;
30
break ;
end
31
32
end
33
34
for k =1: size ( blank (: ,1) )
if tfinish - blank (k ,1) <= 0
35
36
pfinish = k ;
37
break ;
end
38
39
end
40
41
title ( ’ Select emission start and stop of range of interest ’) ;
42
[x , y ] = ginput (2) ;
43
tstart = x (1) ;
104
44
tfinish = x (2) ;
45
46
for k =1: size ( sample (: ,1) )
if tstart - sample (k ,1) <= 0
47
48
estart = k ;
49
break ;
end
50
51
end
52
53
for k =1: size ( sample (: ,1) )
if tfinish - sample (k ,1) <= 0
54
55
efinish = k ;
56
break ;
end
57
58
end
59
60
61
wtdblank = blank (: ,1) .* blank (: ,2) ;
62
wtdsample = sample (: ,1) .* sample (: ,2) ;
63
64
emission = cumtrapz ( sample ( estart : efinish ,1) , wtdsample ( estart : efinish ) ) ;
65
emission = emission ( end ) ;
66
67
blankpump = cumtrapz ( blank ( pstart : pfinish ,1) , wtdblank ( pstart : pfinish ) ) ;
68
blankpump = blankpump ( end ) ;
69
70
samplepump = cumtrapz ( sample ( pstart : pfinish ,1) , wtdsample ( pstart : pfinish ) ) ;
71
samplepump = samplepump ( end ) ;
72
73
yield = 100* emission /( blankpump - samplepump ) ;
74
75
title ( ’ Place Quantum Yield Label ’) ;
76
gtext ( strcat ( ’ PLQY = ’ , num2str ( yield , ’ %6.1 f ’) , ’% ’) ) ; % num2str ( timeconst ) ,’ s ’) ) ;
77
78
title ( strcat ( ’{ PL Quantum Yield of } ’ , material ) ) ;
105
106
List of my Publications
[1] Matthias BAHLKE, Marc BALDO, and Hiroshi MENDOZA. Method of lift-off
patterning thin films in situ employing phase change resists, November 23 2012.
WO Patent 2,012,158,393.
[2] Matthias E Bahlke, Hiroshi A Mendoza, Daniel T Ashall, Allen S Yin, and
Marc A Baldo. Dry lithography of large-area, thin-film organic semiconductors
using frozen co2 resists. Advanced Materials, 24(46):6136–6140, 2012.
[3] Matthias Erhard Bahlke. A novel sublimable mask lift-off method for patterning thin films. Master’s thesis, Massachusetts Institute of Technology, Dept. of
Electrical Engineering and Computer Science, 2011.
[4] Matthias Erhard Bahlke, Marc A Baldo, and Hiroshi Antonio Mendoza. Method
of lift-off patterning thin films in situ employing phase change resists, November 22 2012. US Patent 20,120,295,382.
[5] Daniel N Congreve, Jiye Lee, Nicholas J Thompson, Eric Hontz, Shane R Yost,
Philip D Reusswig, Matthias E Bahlke, Sebastian Reineke, Troy Van Voorhis,
and Marc A Baldo. External quantum efficiency above 100% in a singlet-excitonfission–based organic photovoltaic cell. Science, 340(6130):334–337, 2013.
[6] Marshall P Cox, Hongtao Ma, Matthias E Bahlke, Jonathan H Beck, Theodore H
Schwartz, and Ioannis Kymissis. Led-based optical device for chronic in vivo
cerebral blood volume measurement. Electron Devices, IEEE Transactions on,
57(1):174–177, 2010.
[7] TD Heidel, D Hochbaum, JM Sussman, V Singh, ME Bahlke, I Hiromi,
J Lee, and MA Baldo. Reducing recombination losses in planar organic photovoltaic cells using multiple step charge separation. Journal of Applied Physics,
109(10):104502, 2011.
[8] Jiye Lee, Koen Vandewal, Shane R Yost, Matthias E Bahlke, Ludwig Goris,
Marc A Baldo, Jean V Manca, and Troy Van Voorhis. Charge transfer state
versus hot exciton dissociation in polymer- fullerene blended solar cells. Journal
of the American Chemical Society, 2010.
[9] Hongtao Ma, M Cox, M Bahlke, J Beck, M Zhao, I Kymissis, and T Schwartz.
Novel high frequency surface mounted optical recording system for chronic in107
tracranial optical imaging. In EPILEPSIA, volume 50, pages 393–393. WILEYBLACKWELL PUBLISHING, INC COMMERCE PLACE, 350 MAIN ST,
MALDEN 02148, MA USA, 2009.
[10] Samuel P Subbarao, Matthias E Bahlke, and Ioannis Kymissis. Laboratory
thin-film encapsulation of air-sensitive organic semiconductor devices. Electron
Devices, IEEE Transactions on, 57(1):153–156, 2010.
[11] Nicholas J Thompson, Eric Hontz, Daniel N Congreve, Matthias E Bahlke, Sebastian Reineke, Troy Van Voorhis, and Marc A Baldo. Nanostructured singlet
fission photovoltaics subject to triplet-charge annihilation. Advanced Materials,
2013.
108