© 2014 Jongmin Youn

CARBON NANOSTRUCTURES/TIN (IV) OXIDE SYSTEMS AND

TRANSPARENT ELECTRODE APPLICATION

BY

JONGMIN YOUN

THESIS

Submitted in partial fulfillment of the requirements

for the degree of Master of Science in Materials Science and Engineering

in the Graduate College of the

University of Illinois at Urbana-Champaign, 2014

Urbana, Illinois

Adviser:

Associate Professor Moonsub Shim

ii

ABSTRACT

Interesting electronic properties of sp2 hybridized carbon materials has been attracting an

enormous number of researchers around the globe. Some of the efforts, for more than a decade,

have been in attempting to replace the conventional transparent electrode indium tin oxide (ITO)

with carbon nanostructures, namely carbon nanotube and graphene. In this thesis, carbon

nanostructures and metal oxide hybrid system is studied in an attempt to reduce the resistance of

carbon-based transparent conducting films. It is expected that charge transfer doping in carbon

nanostructures will take place upon contact with metal oxide due to the difference in the work

function. Focusing primarily on single-walled carbon nanotube network, the percolation theory

suggested by Hu and colleagues was used to calculate a figure of merit, DC to optical conductivity

ratio, from optical transmittance and sheet resistance. Among the metal alkoxides tested, tin

isopropoxide was demonstrated to be the best dopant for carbon nanostructures, and carbon

nanotube film and tin (IV) oxide system was studied with electrical measurement and near-IR

absorption and Raman spectroscopies. The concentration of 0.5 to 2% w/v of tin isopropoxide

solution in isopropanol (TIP) was determined to optimize transmittance and conductance of carbon

nanotube and graphene films. A proof-of-concept thin film diode was fabricated with this hybrid

system, which exhibited successful rectifying behavior.

iii

ACKNOWLEDGEMENTS

I would like to deeply thank my thesis adviser, Professor Moonsub Shim, who has been giving me

ideas in research as well as great inspiration as a scientist, engineer and leader. You have helped

me not only to deepen my scientific knowledge but also to improve my problem-solving and

presentation skills, which will be used for the rest of my life. Your unparalleled support has been

the most important factor of my successful graduate school life.

I also want to thank my lab mates in the Shim Research Group for sharing thoughts and teaching

me about research. I thank Johnathon Tsai for helping me settle in this research group and teaching

me how to use the instruments in the laboratory. I am also deeply grateful to Dr. Byung Hwa Seo

for the tremendous amount of discussion we had not only about research but also about life in

general.

iv

TABLE OF CONTENTS

List of Abbreviations ...................................................................................................................... v

CHAPTER 1: INTRODUCTION ................................................................................................... 1

1.1. Graphene .......................................................................................................................... 3

1.2. Carbon nanotube (CNT) ................................................................................................... 3

1.3. Percolation in carbon nanotube network .......................................................................... 6

CHAPTER 2: EXPERIMENTAL METHODS .............................................................................. 8

2.1. Fabrication of carbon nanotube and single-layer graphene films .................................... 8

2.2. Growth of carbon nanotubes via chemical vapor deposition ......................................... 10

2.3. Tin (IV) oxide deposition and characterization ............................................................. 11

2.4. Fabrication of thin-film diode and nanorod heterojunction (NRH) solar cell ............... 13

CHAPTER 3: CARBON NANOTUBE TRANSPARENT CONDUCTING FILM .................... 14

3.1. Surfactant and concentration.......................................................................................... 14

3.2. Choice of metal alkoxide ............................................................................................... 16

3.3. CVD-grown carbon nanotube films ............................................................................... 19

3.4. Carbon nanostructures/Tin(IV) oxide system ................................................................ 22

CHAPTER 4: SINGLE-LAYER GRAPHENE TRANSPARENT CONDUCTING FILM ........ 33

CHAPTER 5: POTENTIAL APPLICATION .............................................................................. 36

CHAPTER 6: CONCLUSION ..................................................................................................... 41

REFERENCES ............................................................................................................................. 43

v

List of Abbreviations

BWF Breit-Wigner-Fano

CNT Carbon nanotube

CVD Chemical vapor deposition

DC/OP ratio DC to optical conductivity ratio

ITO Indium tin oxide

SDBS Sodium dodecylbenzenesulfonate

SLG Single-layer graphene

SWNT Single-walled carbon nanotubes

m-SWNT Metallic single-walled carbon nanotubes

s-SWNT Semiconducting single-walled carbon nanotubes

TCF Transparent conducting film

TCO Transparent conducting oxide

TIP Tin (IV) isopropoxide solution in isopropanol

1

CHAPTER 1

INTRODUCTION

Low-dimensional carbon nanostructures, including carbon nanotubes and graphene, have

been one of the most extensively studied materials around the globe1,2. Carbon-based transparent

electrodes, as a part of application of these materials, have been researched in an effort to substitute

indium tin oxide (ITO), one of the most widely used conventional transparent conductive oxides

(TCOs).3,4

Carbon-based transparent electrodes have different characteristics from ITO. First, the

electronic structure of carbon nanomaterials are so susceptible to the surroundings that it can easily

be changed by adsorption of foreign materials. For example, for the case of organic or quantum

dot optoelectronic devices, ITO is widely used as their transparent anode.5,6 ITO has a relatively

small work function of around 4.6-4.7 eV,7 and since the current density exponentially decreases

with the energy barrier height, organic and/or inorganic hole transport layer (HTL) has been

introduced in order to lower the energy barrier for holes. However, an introduction of additional

layer in between the electrode and the active layer would increase the series resistance of the device.

For the case of carbon-based electrodes, it has been shown that carrier concentration and work

function of carbon nanostructures can be easily modulated not only by an externally applied

electric field,8–10 but also by chemisorption and physisorption of small molecules or polymers.11–

14 Therefore, by p-doping the carbon-based electrode therefore by increasing work function, the

energy barrier for holes can be reduced without introducing additional HTL.

Second, carbon nanostructures are known to be mechanically robust such that devices can

be made on flexible substrates.15–17 Resistances of both carbon nanotube network and graphene

film does not significantly change upon bending unlike ITO where resistance change can be as

2

high as 70 times the original resistance with the strain of 0.02,18 and researchers have successfully

constructed single and few-layer graphene-based touch screen on flexible substrates.19 Further, for

the case of carbon nanotube network, surface structure is rough such that the current per nominal

area can be increased. The thickness of carbon nanotube network typically ranges between 5 and

50 nm,20 and there are still pinholes in the film down to the substrate. If this network can be

conformally coated, surface roughness is advantageous for organic and quantum dot optoelectronic

devices in that the practical contact area among the electrode, carrier transport layers and the active

layer can be very large.

However, the most substantial issue in carbon-based electrode is that its resistance is

significantly higher than that of ITO. Carbon nanotube network has roughly 200 Ω/□ of sheet

resistance with roughly 90% transmittance in visible spectrum range, whereas well-documented

and widely used ITO has sheet resistance of 10-100 at 97% transmittance.21 Large-area single-

layer graphene (SLG) has a comparable transmittance of 97.5%19 against ITO, and its sheet

resistance is on the order of 1,000 Ω/□.

This high resistance can be reduced by non-covalently charge transfer doping. Researchers

around the globe have shown for a decade that not only sheet resistance reduction but also work

function modulation of carbon nanotube network and graphene are possible by treating them with

nitric acid,11,19,22,23 thionyl chloride,24 gold chloride25–27 or other small molecules that attracts

charge carriers.12,28 Some researchers utilized carbon nanotube network and graphene hybrid film

to modulate work function of their transparent electrodes.29 We deposited tin (IV) oxide by

hydrolysis and condensation of metal alkoxide for charge transfer doping of carbon nanostructures.

Due to the offset in Fermi level positions between carbon nanomaterials and tin (IV) oxide,

electrons will be trapped at the localized state of tin (IV) oxide, and this results in net increase of

3

delocalized holes in the carbon nanostructures. This also leads to shifting in the Fermi level and

reduction in resistance.

Here, charge transfer in carbon nanostructures/tin (IV) oxide thin film system is studied,

focusing on improvement and application of carbon-based transparent electrode. Not only their

sheet resistances and transmittance in visible wavelength but also near-IR absorption and Raman

spectra were systematically analyzed as functions of the concentration of the tin (IV) oxide

precursor, tin isopropoxide solution in isopropanol (TIP). A proof-of-concept thin film diode based

on this system has also been demonstrated.

1.1. Graphene

Graphene is one of the most extensively studied low-dimensional materials, and interesting

optical and electronic properties and phenomena have been theoretically and experimentally

discovered, such as near-ballistic carrier transport with exceptionally high mobility and massless

relativistic Dirac fermion.2,30 Single- and few-layer graphene have been studied as channel

materials for field-effect transistors utilizing its zero-band-gap nature,31–33 and as transparent

electrodes for optoelectronic devices thanks to its high electrical conductivity with atomic

thickness.19,34 In this work, sheet resistances of large-area SLG and SLG/tin (IV) oxide system

were measured by the four-point probe method in an effort for its application in transparent

electrode.

1.2. Carbon nanotube (CNT)

Geometrically, a single-walled carbon nanotube (SWNT) has a form of one layer of

graphene sheet rolled up seamlessly to form a hollow rod composed of sp2 hybridized carbon

4

atoms.35 Thanks to its one-dimensional electron cloud, carbon nanotubes exhibit interesting

properties such as exceptionally high mobility with ballistic carrier transport36 and other

phenomena due to two-dimensional quantum confinement.37

There are two types of single-walled carbon nanotubes depending on the presence of band

gap. A semiconducting single-walled carbon nanotube (s-SWNT) has a band gap typically around

0.5-0.7 eV, and a metallic single-walled carbon nanotube (m-SWNT) has a zero band gap.35

Statistically one third of single-walled carbon nanotubes are metallic and the other two thirds are

semiconducting.35 In this study, a random mixture of arc-discharge single-walled carbon nanotubes

was used, and each type of carbon nanotubes was characterized by near-Infrared (near-IR)

absorption and Raman spectroscopies.

The charge transfer doping in single-walled carbon nanotubes can be observed by near-IR

absorption spectra. The density of states of single-walled carbon nanotubes is extensively studied

and well-documented.38,39 The density of states spikes at certain energy levels and exponentially

decreases until the next spike. These spikes are known as van Hove singularities, which originate

from quantum confinement in two dimensions. The optical absorption originate from electronic

transitions from one of the spikes in the valence band to the corresponding one in the conduction

band.39,40 These optical transitions, which results in absorbance peaks in near-IR absorption spectra,

are often labeled as S11, S22 and M11 for the first and the second spikes of semiconducting carbon

nanotubes, and the first spike for metallic nanotubes, respectively.38

A shift in the Fermi level position in carbon nanotubes results in a bleach in the S11 optical

transition mode, or attenuation of the absorbance peak of the lowest energy in near-IR absorption

spectra, and this is often used to indicate the doping level in nanotubes.39,41 In this study, near-IR

5

absorption spectra of carbon nanotube network was systematically analyzed to show the doping

level change in semiconducting nanotubes.

The Raman spectroscopy has been one of the most widely used characterization methods

for sp2 hybridized carbon allotropes including carbon nanotubes and graphene. It is a non-

destructive characterization which enables observation of doping level, defects, strain, and

functionalization.42,43 While sp2 hybridized carbon materials have a number of different optical

and acoustic in-plane and out-of-plane phonon modes as explained by Dresselhaus and

colleagues,35 a Raman spectrum of a single-walled carbon nanotube has four main distinct features.

The radial breathing mode (RBM) at the frequency range between 100 and 300 cm-1 represents the

out-of-plane lattice vibration, which can also indicate the diameter of single-walled carbon

nanotubes. The defect mode at around 1320 cm-1, also known as the D band, comes from phonon

scattered by defects in the honeycomb carbon sheet, which also can be indicative of the degree of

functionalization.44 The tangential mode that ranges from 1550 to 1610 cm-1, widely referred to as

the G band, is the in-plane vibration of sp2 hybridized carbon sheet, and this mode is seen not only

in carbon nanotubes and graphene but also in amorphous carbon.33,35,45 Lastly, the two-phonon

mode or the 2D band around 2630 cm-1 is from the second-order scattering of phonons.35,42,43

The lineshapes of the tangential mode in metallic and semiconducting carbon nanotubes

are different. For the case of metallic carbon nanotubes, the G band includes the form of

asymmetric Breit-Wigner-Fano (BWF) lineshape, which is due to the presence of zero-band-gap

point in the momentum space and resulting non-adiabatic Kohn anomaly.9,42,46 In contrast, that of

semiconducting nanotubes is a single Lorenzian peak due to non-zero band gap.45 Furthermore,

the asymmetry of the BWF peak or the position of the Lorenzian peak is highly dependent upon

gate voltage9,10 or adsorbate,13,47 thus carrier concentration and Fermi level in the carbon nanotube.

6

As the Fermi level is shifted to either side from the charge neutrality point, the BWF lineshape not

only blue shifts but also gradually changes into a Lorenzian lineshape (or the Lorenzian peak is

blue-shifted for semiconducting SWNTs). This is caused by electron-phonon coupling, and the

spectral change is often used to indicate charge transfer doping in carbon nanotubes.47 The gate

voltage dependence in the BWF lineshape in a single metallic carbon nanotube is systematically

studied and well-illustrated by Nguyen and colleagues.9

In this research, a random mixture of arc-discharge single-walled carbon nanotubes was

used. Therefore, the RBM of carbon nanotube network contains a number of peaks ranging from

150 to 200 cm-1. For the same reason, the G band becomes a mixture of BWF and Lorenzian

lineshapes. Primarily, Raman spectra of carbon nanotube network was systematically analyzed to

study the change in intensity and peak position in G band in relation to the doping level, focusing

on metallic nanotubes.

1.3. Percolation in carbon nanotube network

Figure 1.11 shows a typical surface morphology of a carbon nanotube network, in which

dense, random network of carbon nanotubes and occasional metal catalyst particles are visible.

While there is a two-dimensional electron gas in a metallic thin film, a carbon nanotube is a one-

dimensional carbon allotrope, so the conduction in carbon nanotube network takes place by charge

carriers hopping from one nanotube to another through percolation pathways. The percolation

model in carbon nanotube approximates that each nanotube is a conductive rod and these rods are

randomly oriented to form a rough two-dimensional plane. The sheet resistance and the amount of

light that goes through the plane is related by the following20,48

𝑅𝑆 =188.5

(𝑇−12⁄ − 1) ∙

𝜎𝐷𝐶𝜎𝑂𝑃

⁄ 𝐸𝑞𝑛 1

7

where T is transmittance at 550 nm in range of 0 to 1, RS is sheet resistance in Ω/□ and σDC and

σOP are DC and optical conductivity, respectively. 𝜎𝐷𝐶

𝜎𝑂𝑃⁄ is also commonly known as the DC to

optical conductivity ratio (DC/OP ratio), which is a unitless fitting parameter in the sheet resistance

versus transmittance curve.

In order for this approximation to work, first, the film thickness should be less than the

wavelength of light that was measured, and second, the amount of reflected light by the film should

be smaller than that of the absorbed light. This relation can be well applied to carbon nanotube

films, because it was shown that the typical thickness of carbon nanotube network ranges from 5

to 50 nm whose transmittance ranges from 60% to 99%,20 and the reflection by carbon nanotube

films was almost ten times less than the absorption.49 Improvement of network film conductivity

involves the sheet resistance and transmittance curve shifted down so that the film exhibits lower

sheet resistance at a fixed transmittance.

Figure 1.1: Single-walled carbon nanotube network on a Si/SiO2 wafer.

Metal catalyst particles are also occasionally present.

8

CHAPTER 2

EXPERIMENTAL METHODS

2.1. Fabrication of carbon nanotube and single-layer graphene films

Arc-discharge single-walled carbon nanotube (SWNT) powder and sodium

dodecylbenzene sulfonate (SDBS) were purchased from Carbon Solutions (AP-SWNT) and Sigma

Aldrich, respectively. SWNT powder of 1 g/L and SDBS of 0.1 wt% were mixed with deionized

water. The solution was homogenized with an 80 W bath sonicator (Fisher Scientific) for 30

minutes and with a horn sonicator (Fisher Scientific Model 505) at 250 W for 1 hour. The power,

50% amplitude of a 500-watt sonicator, was applied with 1-second interval, so the overall

sonication time was 2 hours. The solution was transferred into 1.5-ml micro-centifuge tubes and

then centrifuged (Revolutionary Science RS-200) at 10,000 rpm (roughly 6,000G) for 1 hour. Only

the clear SWNT dispersion on top of the centrifuge tubes was transferred and used. Care should

be taken since the sediment may be swept into the target container while transferring the

supernatant. The SWNT dispersion was sprayed (Grex Tritium.TS) with a stream of lab grade

nitrogen gas onto half-inch-by-half-inch square glass substrates that were heated up to 110°C. The

distance from the nozzle and the substrate was fixed at around 20 cm and the amount of SWNT

solution was adjusted in order to control the resistance of the SWNT film. Spray-deposited SWNT

films were cooled for 2 minutes and immersed into deionized water for 10 minutes to wash out the

residual surfactant, and dried on a hot plate at 60°C for 10 minutes. SWNT films soaked with

deionized water is fragile, so the films should be dried carefully, slowly and thoroughly.

Arc-discharge SWNT was chosen, because for CVD-grown SWNT or double-walled

nanotube (DWNT) powder, the concentration of CNT dispersion was low and sprayed CNT films

9

were not stable enough so that they easily ripped when washed and dried. More importantly, their

sheet resistances were orders of magnitude higher than those of arc-discharge SWNTs.

In order to test the surfactant-dependent performance of carbon nanotube network, other

anionic, cationic and nonionic surfactants, such as 0.5 wt% sodium dodecylsulfonate (SDS), 1 wt%

sodium cholate (Na Cholate), 0.1 wt% cetrimonium bromide (CTAB), 0.1 wt% Triton X-100 and

0.1 wt% Tween-20, were also used. These concentrations for each surfactant were chosen so that

they are above the critical micelle concentration (CMC). Further, the surfactant-concentration-

dependent performance of carbon nanotube network was tested by changing the concentration of

SDBS to 0.1, 0.5 and 1.0 wt%. The rest of the processing steps from sonication to spraying onto

substrates are the same as above. For all the carbon nanotube films tested, 1 g/L SWNT powder

and 0.1 wt% SDBS were consistently used unless indicated.

Single-layer graphene was grown on Cu foil by chemical vapor deposition (CVD) and was

transferred onto a designated substrate, either a Si/SiO2 wafer or a glass slide, following the

identical process as the one by Baeumer et al.50 To briefly summarize, after high-purity Cu foil

(99.999%, Alfa Aesar) was annealed for 30 minutes at 1,000 °C under hydrogen atmosphere with

the pressure around 300 mtorr, single-layer graphene was grown by flowing both 7 sccm of

hydrogen and 60 sccm of CH4 for 20 minutes. The system was rapidly cooled using a fan while

both gases were still flowing, and when the temperature reaches roughly 150 °C, gas supply was

stopped and SLG on Cu foil was transferred out. Then the SLG film on Cu foil was covered by

thin poly(methyl methacrylate) (PMMA) film (A2 495K, 2% solids in Anisole, Microchem) by

spin-coating the solution at 3,000 rpm for 30 seconds, and the SLG on the other side of the Cu foil

was removed by reactive ion etching. The SLG/PMMA/Cu film was placed on the surface of 0.1

M aqueous ammonium persulfate (Sigma Aldrich) solution. After the Cu foil completely dissolved,

10

the SLG/PMMA film was transferred onto deionized water with a glass slide to wash out residual

ammonium persulfate for about an hour, and it was transferred onto the designated substrate. Once

deionized water between the substrate and the film completely dried in the ambient condition, the

substrate was immersed in acetone for an hour to remove PMMA film.

2.2. Growth of carbon nanotubes via chemical vapor deposition

Single-walled carbon nanotubes were grown by catalytic chemical vapor deposition

(CCVD).51 Fe/Mo/Al2O3 catalyst was prepared by mixing 52.3 mg of Fe(NO3)3·9H2O (Aldrich),

3.1 mg of MoO2(acac)2 (Aldrich), 30.0 mg of Al2O3 powder and 10 ml of methanol. The solution

was spin-coated on quartz substrate at 2,000 rpm for 30 seconds. The coated quartz substrate was

transferred into a 1-inch-diameter quartz tube in a furnace. Hydrogen, argon and methane gases

were supplied into the tube with flow rates of 150, 150 and 300 sccm, respectively. Carbon

nanotubes were grown at 900°C for 10 minutes. Multiple quartz substrates with CVD CNTs were

submerged in 30 ml of deionized water with 0.5 wt% SDBS, sonicated and centrifuged as

previously discussed in an attempt to disperse CVD CNTs in water. The aqueous solution was

sprayed on glass or SiO2 substrates in the same manner.

Single-walled carbon nanotubes were also grown by floating catalyst chemical vapor

deposition (FCCVD).52 Ferrocene powder (Aldrich) was mixed with 0.5 wt% sulfur powder. The

mixture was transferred to the reactor tube and was sublimated at 75°C during growth. The

sublimation was carried to the reaction spot of 1,100°C by 2,000 sccm of hydrogen and 1,000 sccm

of argon. Less than 5 sccm of methane was used as the carbon source, and carbon nanotubes were

grown for 20 minutes. Carbon nanotube bundles and ropes were carefully retained from the water-

11

cooled cold finger and the wall of the CVD tube. The resulting powder was dispersed in water

with 0.5 wt% SDBS, and the solution was sprayed onto substrates in the same method above.

2.3. Tin (IV) oxide deposition and characterization

Tin (IV) oxide was deposited by hydrolysis and condensation reaction of tin alkoxide. Tin

(IV) isopropoxide solution in isopropanol (10% w/v, Alfa Aesar) (TIP), a well-known precursor

of tin (IV) oxide, was directly spin-coated on a SWNT or SLG film at 2,000 rpm for 30 seconds.

Hydrolysis and condensation takes place when the solution is spin-coated in ambient condition

reacting with water vapor. Then the film was immersed in deionized water for 5 minutes to

completely hydrolyze the isopropoxide, blow-dried with a stream of nitrogen gas, and dried on a

hot place at 60 °C for 5 minutes. The concentration, which is 10% w/v as-purchased, was

controlled by adding sufficient amount of anhydrous isopropanol (Sigma Aldrich) in a glove box.

Other metal alkoxide were also used in order to compare their ability to improve the

performance of SWNT films. Zinc tert-butoxide powder, molybdenum (V) isopropoxide powder

and powder mixture of zinc tert-butoxide and tin (IV) isopropoxide (Alfa Aesar) were used as

precursors for zinc oxide (ZnOx), molybdenum (V) oxide (MoOx) and zinc oxide: tin oxide

(ZnOx:SnOx) respectively. Stoichiometric factors of metal oxides were not characterized; therefore,

they are denoted as ‘Ox’ throughout this paper. Zinc tert-butoxide and molybdenum isopropoxide

solutions of 0.42% w/v and 0.78% w/v in isopropanol were spin-coated on SWNT films at 1,500

rpm for 30 seconds in ambient condition. Further, 1:1 molar-ratio mixture of zinc-tert-butoxide

and tin isopropoxide solution was prepared by mixing 0.42% w/v zinc tert-butoxide and 0.71%

w/v tin isopropoxide solutions in isopropanol, and was also spin-coated on SWNT films with the

12

same spin-coating recipe. All of the concentrations used in this particular step is equivalent to 20

mM.

SWNT solution was sprayed onto silicon and silicon dioxide wafer square pieces and were

spin-coated by 0.71% w/v TIP in isopropanol at 1,500 rpm for 30 seconds in order to observe the

surface morphology with a scanning electron microscope (Hitachi S-4800). SWNT solution was

also sprayed on 400-mesh Cu grids with holey carbon support film (Ted Pella) and spin-coated by

TIP solution in the same manner to obtain a magnified view of SWNT/tin (IV) oxide system with

a transmission electron microscope (JEOL 2100).

Silver conductive epoxy (MG Chemicals) was applied onto four corners of a square SWNT

or graphene film for Van der Pauw measurement of sheet resistance (Agilent Semiconductor

Analyzer 4156C). Current was swept through two adjacent corners from -100 μA to 100 μA, and

voltage across the other two adjacent corners was measured. Sheet resistance, RS, was calculated

from the following analytic expression:

𝑒−𝜋

𝑅12,34𝑅𝑆

⁄+ 𝑒

−𝜋𝑅23,41

𝑅𝑠⁄

= 1 𝐸𝑞𝑛 2

where R12,34 refers to the resistance in Ω derived when current flowed through the first and the

second corners and voltage was measured across the third and the fourth corners, and R23,41 is the

resistance when current flowed through the second and the third, and voltage was across the fourth

and the first.

Transmittance spectrum was obtained (Ocean Optics HR2000), and the transmittance at

550 nm wavelength was recorded for potential application purposes (source). Near IR spectra of

SWNT network was taken between 4,000 cm-1 and 11,000 cm-1 with an InGaAs detector (Thermo

Nicolet Nexus 670). Raman spectra of SWNT films before and after deposition of tin oxide were

obtained between 100 cm-1 and 3000 cm-1 with 633 nm laser (JY LabRAM HR800). Spectra were

13

taken at 6 different spots for each film, were normalized and averaged, and the peaks between

1450 cm-1 and 1650 cm-1 were primarily analyzed.

2.4. Fabrication of thin-film diode and nanorod heterojunction (NRH) solar cell

Solar cells with carbon nanostructure transparent electrode with and without NRH active

layer were fabricated. The NRH was synthesized by following the method of McDaniel et al.53

The NRH was recapped in situ with octanethiol for enhanced stability. First, SWNT films were

prepared and 10% w/v tin isopropoxide was spin-coated. Tin dioxide powder (Sigma Aldrich) was

mixed with isopropanol in 3% w/v and was also spin-coated twice at 2,000 rpm for 30 seconds. A

poly(3-octylthiophene-2,5-diyl) (P3OT) solution was prepared by mixing 34 mg of P3OT with

0.29g of chloroform and 12.6 g of xylenes.53 The NRH and P3OT solutions were spin-coated onto

the SWNT/SnOx/SnO2 film sequentially, at 2,000 rpm for 30 seconds for both solutions, and the

device was baked at 100 °C under ~100 torr vacuum environment for 24 hours. A ~50 nm Au

contact was deposited with an electron beam evaporator on a separate glass substrate, and the

device and the contact substrates were put together with a binder clip.

14

CHAPTER 3

CARBON NANOTUBE TRANSPARENT CONDUCTING FILM

3.1. Surfactant and concentration

Figure 3.1 presents a sheet resistance versus percent transmittance at 550 nm for different

surfactants used, and DC/OP ratios for these sets of samples are analyzed and tabulated in Table

3.1. As stated in the experimental section, anionic, cationic and nonionic surfactants, namely

SDBS, SDS, Na Cholate, CTAB, Triton X-100 and Tween-20 were tested. The SWNT films with

SDBS and SDS as dispersants are occupying the lowest part of the plot in Figure 3.1, meaning

their sheet resistances are the lowest at the same transmittance compared to those samples with

other surfactants. DC/OP ratios of these SWNT films show that 1) ionic surfactants yield better

conductivity than nonionic surfactants do, 2) linear surfactants—SDBS, SDS and CTAB—exhibit

better conductivity than bulky surfactants do. From the figures of merit derived from sheet

75 80 85 90 95 10010

2

103

104

105

0.5 wt% SDBS

0.5 wt% SDS

0.1 wt% Triton X-100

0.1 wt% Tween-20

0.1 wt% CTAB

1 wt% Na Cholate

Sh

ee

t R

esis

tan

ce

(

/sq

)

Transmittance at 550 nm (%)

Figure 3.1: Sheet resistance of carbon nanotube network as a

function of optical transmittance with different surfactants used.

Dotted and dashed trend lines represent fitting curves for

samples with SDS and SDBS, respectively.

15

resistances and transmittances of SWNT samples, SDBS was selected for the best dispersant for

SWNT aqueous solution.

Concentration-dependent sheet resistance and transmittance of SWNT film was also tested

as shown in Figure 3.2, and Table 3.2 shows DC/OP ratios of SWNT films with three different

SDBS concentrations. When the transmittance of SWNT film is high around ~95% where

reduction of transmittance is dominated by SWNTs, sheet resistances of films are in a similar range;

however, when the transmittance becomes lower, reduction in sheet resistance is higher for the

films with lower SDBS concentration. It is thought that as the film becomes thicker, more

surfactant molecules are trapped in the SWNT network, contributing to not only reduction in

transmittance but also steric hindrance between SWNTs, increasing barriers for charges to

percolate through the network. The lowest SDBS concentration among three, 0.1 wt%, was

selected since it yielded the highest DC/OP ratio and it is above the CMC of SDBS, which is

known to be 1.5 mM or 0.05 wt% in water.54

Table 3.1: DC/OP ratio of carbon nanotube network for different surfactants

used to disperse nanotubes.

16

3.2. Choice of metal alkoxide

Figure 3.3 exhibits the result of deposition of different metal alkoxide solutions, and Table

3.3 shows DC/OP ratios of SWNT films before and after deposition of the solutions. Four sets of

four SWNT films that have similar ranges of sheet resistance and transmittance were used for each

20 mM metal alkoxide solution in an effort to deposit different metal oxides: SnOx, ZnOx, MoOx

75 80 85 90 95 100100

1000

10000

0.1 wt% SDBS

0.5 wt% SDBS

1.0 wt% SDBS

Sheet R

esis

tance (

/sq)

Transmittance at 550 nm (%)

Figure 3.2: Sheet resistance of nanotube film versus optical

transmittance with different concentration of SDBS used. Trend

lines are fitting curves based on Eqn 2.

Table 3.2: DC/OP ratio of nanotube network for

different SDBS concentration used.

17

70 80 90 100100

1000

SnOx

ZnOx

MoOx

ZnOx:SnOx

Sheet R

esis

tance (

/sq)

Transmittance at 550nm (%)

Figure 3.3: Sheet resistance of nanotube network versus

transmittance for different metal oxides for dopants. Dotted and

dashed trend lines are for samples with ZnOx:SnOx and SnOx,

respectively.

Table 3.3: DC/OP ratio of nanotube network with

different metal oxide dopants.

18

and ZnOx:SnOx. Since each set of films had similar sheet resistance and transmittance before each

treatment, it is shown in Figure 3.3 that deposition of SnOx did not reduce transmittance as much

as other metal oxide deposition did. Transmittance of ZnOx-deposited SWNT films was

approximately the same as the ones with SnOx; however, deposition of MoOx reduced the

transmittance by ~2-3%, and that of ZnOx:SnOx reduced by ~10%. As shown in Table 3.3, DC/OP

ratio for SnOx-deposited films is the highest among all the metal oxide precursors used, and this is

represented by their trend line in the sheet resistance vs transmittance curves located at the lowest

part of Figure 3.3.

Sheet resistances of SWNT films with different metal oxides was plotted as a function of

bulk values of the metal oxide work functions5,55–58 along with as-prepared SWNT film as shown

in Figure 3.4, and the data points coincide in a linear curve with the exception of MoOx-deposited

sample. Even though the figure suggests possible scalability of work function of SWNT film with

4.0 4.5 5.0 5.5 6.0 6.5 7.0

400

600

800

1000

1200

1400

As-prepared

SnOx

ZnOx

MoOx

ZnOx:SnOx

Sh

ee

t R

esis

tan

ce

(

/sq

)

Work Function (eV)

Figure 3.4: Sheet resistance of nanotube network with different

metal oxides as dopants as a function of bulk value of work

function of metal oxides. The dashed trend line is a fitted line for

four data points excluding MoOx deposited nanotube network.

19

different metal oxides, further studies are required on the optical and electronic properties of metal

oxides formed by hydrolysis and condensation, especially their work functions. Even though the

work function of bulk MoO3 is known to be around 6.7 eV,56 it is learned that the work function

of molybdenum oxide changes with respect to its stoichiometry, ranging from 5.6 eV to 6.8 eV.59

Moreover, it is reported that low-dimensional MoO3 systems including thin film or nanoparticles

has lower work function than that of bulk.60 Therefore, the electronic properties of not only MoO3

but also other metal oxide nanoparticles should be studied to further understand dopant work

function dependence of carbon nanotube film.

3.3. CVD-grown carbon nanotube films

Single-walled carbon nanotubes were grown by CVD with Fe/Mo/Al2O3 catalyst in order

to have high-quality carbon nanotubes for transparent conducting films. Carbon nanotubes were

grown on quartz substrates, and the Raman spectra of the nanotubes are presented in Figure 3.5.

Each spectrum was obtained from different substrates in different CVD growths with the same

growth conditions as indicated in the CHAPTER 2 . Not only the change in the quality of carbon

Figure 3.5: Raman spectra of carbon nanotubes grown by Fe/Mo/Al2O3 catalytic CVD. a. Full

Raman spectra which includes RBM, D, G and 2D bands. b. G band spectra of the carbon

nanotubes.

20

nanotubes, represented by the height ratio between D and G band, and the shape of G band but

also irregular peak positions in RBM exhibit inferior reproducibility of this carbon nanotubes.

Carbon nanotubes were grown in multiple quartz substrates, dispersed in water with SDBS,

and sprayed on glass and Si/SiO2 substrates in the same way as the arc-discharge carbon nanotube

dispersion. The solution was observed to be dark enough as shown in Figure 3.6 a, meaning that

it might contain a sufficient amount of carbonaceous material. However, the optical density of this

dispersion was too low in near-IR regime, and there was no observable S11 transition peak in the

near-IR absorption spectrum as indicated in Figure 3.6 b. Further, carbon nanotubes were not seen

with SEM. Van der Pauw measurement was attempted and the circuit was open, meaning that

percolation pathway from one contact to another was not formed. It is concluded that Fe/Mo/Al2O3

catalytic CVD is not suitable for a transparent conductor application.

Floating catalyst CVD (FCCVD) was also explored to get a large amount of carbon

nanotube bundle. FCCVD is known for capability of grow diameter-specific carbon nanotubes by

controlling gas flow rates and growth temperature.52 Parameters were chosen to grow large-

diameter m-SWNTs. CNT bundle and rope were obtained not only on the surface of cold finger

Figure 3.6: a. Direct comparison of CNT dispersions grown by arc-discharge and Fe/Mo/Al2O3

CCVD methods. b. Near-IR absorption of arc-discharge CNT dispersion (black line) and CCVD

CNT dispersion (red line).

21

but also on the wall of the reaction tube, and the Raman spectra are shown in Figure 3.7. Peaks in

spectra are well-defined and the D/G ratio is small; however, the peak positions in RBM does not

seem to be consistent.

FCCVD-grown carbon nanotubes were dispersed in water with SDBS and sprayed on glass

slides in a similar manner. Transmittance, sheet resistance and calculated DC/OP ratio based on

this carbon nanotubes are presented in Table 3.4, and those based on arc-discharge CNTs are

included as a comparison. The transmittance of FCCVD-grown CNT TCF is lower than that of

arc-discharge CNT TCF, and the sheet resistance is larger by more than a factor of five. In turn,

DC/OP ratio of FCCVD-grown CNT TCF is an order of magnitude smaller. This might not readily

mean that FCCVD-grown CNTs are inferior to arc-discharge CNTs. It is because the FCCVD

growth was not optimized as seen from irregular peak positions in RBM. Further, those nanotubes

Figure 3.7: Raman spectra of carbon nanotubes grown by FCCVD. a. Full spectra in which RBM,

D, G and 2D band are present. b. The G band of FCCVD CNTs.

Table 3.4: Optical transmittance, sheet resistance and DC/OP ratio of transparent conducting films

based on FCCVD CNTs and arc-discharge CNTs.

22

did not go through the purification process; therefore, they might contain amorphous carbon and

large catalyst particles between them, which would cause steric hindrance in percolation pathway

of carbon nanotubes. Further research effort is required to understand and solve this issue.

3.4. Carbon nanostructures/Tin(IV) oxide system

Figure 3.8 shows scanning and transmission electron microscopy images of spray-

deposited carbon nanotube network. Figure 3.8 a and c are the network before TIP deposition and

Figure 3.8 b and d are the ones after spin-coating TIP. It is seen that tin (IV) oxide does not form

a uniform film; rather it forms nanoparticles on the sidewall of carbon nanotubes. Tin (IV) oxide

does not form a continuous film, and this appears to be because the surface of carbon nanotube

network is rough, and the hydrolysis and condensation reaction during spin-coating is relatively

fast. The nanotube network typically has a thickness around 5 to 50 nm20 and it still has pinholes

through the substrate, and this roughness is thought to inhibit conformal coating of metal oxide

film.

A sheet resistance vs transmittance plot of carbon nanotube film is shown in Figure 3.9.

Each square, triangle or solid circle represent each set of sample, and fitting curves are drawn

based on the previously discussed theory proposed by Hu and colleagues about percolation in

nanotube network.48 The dotted trend line for as-sprayed nanotube film is located on the top of the

plot, and with 0.5% w/v TIP spin-coated, the curve is shifted downwards. A deposition of 10%

w/v TIP also shifts the curve downwards as shown. However, the magnitude of the shift is

seemingly close to 0.5% w/v TIP deposition for the samples of ~85% transmittance, and for the

samples of ~92% transmittance, the data points are on the fitting curve of as-sprayed pristine

23

nanotube films. This offset is further analyzed by calculating DC/OP ratio for each sample, which

is shown Figure 3.10 a.

DC/OP ratio of samples are calculated and plotted as a function of tin isopropoxide

concentration used. Samples that have around 90% transmittance and 80% transmittance are

separately plotted. As soon as we deposit a small amount of TIP around 0.5% w/v, DC/OP ratio is

increased about 25 to 30%, and this means a reduction of sheet resistance in the same magnitude

for the films having the same transmittance. For the case of high-transmittance samples, as we

Figure 3.8: Electron microscopy images of carbon nanotube networks. a,b: scanning electron

microscopy images of nanotube networks before and after deposition of SnOx, respectively. c,d:

transmission electron microscopy images of nanotube networks before and after SnOx deposition.

24

80 85 90 95100

1000

Undoped

10% w/v

0.5% w/v

Sh

ee

t R

esis

tan

ce

(

/sq

)

Transmittance at 550 nm (%)

Figure 3.9: Sheet resistance versus transmittance plot of as-

prepared (open black square), 10% w/v SnOx-deposited (solid

red circle) and 0.5% w/v SnOx-deposited (solid blue triangle)

nanotube films. Dashed trend lines are based on the percolation

in nanotube network.

Figure 3.10: a. DC/OP ratio as a function of TIP concentration for the nanotube films with around

90% transmittance (black square) and around 80% transmittance (red circle). b. Decrease in sheet

resistance (open black square) and decrease in transmittance (open blue circle) as functions of TIP

concentration.

25

increase the concentration of TIP, the DC/OP ratio does not change significantly until 3% w/v, and

it decreases to the same level as pristine carbon nanotube films when the concentration reaches

10% w/v. This corresponds to the data points of nanotube films with 10% w/v TIP located on the

trend line for the undoped pristine nanotube films. On the other hand, for relatively low-

transmittance samples with around 80% transmittance, the DC/OP ratio is increased again with

only 0.5% w/v of TIP. The increase in concentration of TIP does not change the DC/OP ratio. This

is thought to be the case because the reduction in transmittance for the high-transmittance nanotube

films is dominated by tin (IV) oxide nanoparticles. As TIP concentration is increased, more tin

(IV) oxide particles physisorb, and this results in not only decreased sheet resistance but also

decreased transmittance. For the low-transmittance nanotube films, the dominating factor of

transmittance decrease is the nanotube network itself; therefore, increasing TIP concentration

partially contributes to transmittance reduction, but the DC/OP ratio is fairly stationary. This is

precisely what Figure 3.10 b exhibits.

Figure 3.10 b shows that sheet resistance is reduced by roughly 40% when TIP is deposited

and the reduction level is saturated. However, the transmittance decrease is a relatively weaker

function of TIP concentration. This indicates that a deposition of minimal concentration of TIP

largely reduces resistance while there is no significant drop in optical transmittance, and a

deposition of high-concentration TIP is not desirable in the perspective of both resistance and

transmittance. Sheet resistances of nanotube films seem to drastically decrease as TIP

concentration slightly increases, and the magnitude of resistance decreases attenuated for the case

of higher TIP concentration used. This is thought to be related to the size and resulting electronic

property change in tin (IV) oxide nanoparticles, which will be discussed later in this chapter.

26

The carbon nanotube network and tin (IV) oxide particle hybrid system has been studied

not only by optical and electrical measurements, but also with near-IR absorption and Raman

spectroscopies. Figure 3.11 exhibit some of near-IR absorption spectra of nanotube networks

before and after deposition of SnOx. It is seen that the S11 peak located around 6,000 cm-1 is

bleached with only 0.25% w/v of TIP, indicating charge-transfer doping in carbon nanotubes.

Increasing TIP concentration still bleaches the S11 peak; however, the reduction in peak height is

mitigated. The change in peak height for S11 and S22 transitions, located around 6,000 cm-1 and

10,000 cm-1 respectively, are plotted as functions of TIP concentration in Figure 3.12.

Figure 3.11: near-IR spectra of carbon nanotube network before (black) and after (red) SnOx

deposition of respective concentration. a. 0.25% w/v, b. 0.5% w/v, c. 3% w/v and d. 5% w/v.

27

The peak height decrease in S11 transition reaches about 25% with only 0.25% w/v of TIP.

As the concentration is increased, the bleach in S11 peak is maintained up to 2% w/v TIP and is

reduced with higher concentration. This plateau at around 0.5 to 2% w/v TIP coincides with the

result of electrical measurement and DC/OP ratio shown in Figure 3.10. This range is thought to

be the TIP concentration that maximizes the charge transfer doping and sheet resistance reduction

in carbon nanotube networks while minimizing the transmittance reduction. The peak height

change in the S22 transition is also included in this plot, showing decrease in peak height over 2%

w/v TIP. However, the absorbance in S22 is known to be less susceptible to Fermi level change,

because the second van Hove singularity is farther away from the energy midgap of

semiconducting nanotubes.39 Furthermore, the wavenumber range for the InGaAs detector used in

this study was between 4,000 and 10,000 cm-1, where the S22 peak position was slightly higher

than 10,000 cm-1, which further undermines the reliability. Therefore, a setup with a proper

detector should be used in order to study the concentration-dependent bleach in the S22 transition.

Figure 3.12: Percent decrease in peak heights corresponding S11

(open black square) and S22 (open blue circle) optical transitions

as functions of TIP concentration.

0 2 4 6 8 10

0

10

20

30

Decre

ase in S

11 (

%)

Tin Isopropoxide Concentration (% w/v)

0

10

20

30

Decre

ase in S

22 (

%)

28

Where near-IR absorption spectra indicates charge transfer doping in s-SWNTs, the G band of

Raman spectra was analyzed to prove doping level change in m-SWNTs.

While obtaining Raman signal from carbon nanotube films, the laser spot size, which is

roughly 1 µm diameter, was larger by orders of magnitude than the diameter of single carbon

nanotube. It is expected that a number of SWNTs within tens-of-nanometer-thick nanotube

network will be excited by the 1-um-diameter laser, and spectra were taken from six different spots

on one sample. Therefore, Raman spectra in this study are mixtures of spectra from m-SWNTs

and s-SWNTs. Figure 3.13 and Figure 3.14 a show normalized Raman spectra of nanotube

networks with different TIP concentration deposited. As seen in Figure 3.13 a, there is no strong

correlation between the intensity and TIP concentration in any peaks in radial breathing mode.

This means the out-of-plane frequency of vibration in SWNTs are not changed, so deposited tin

oxide nanoparticles are adsorbed physically, not chemically. Further, there is also no correlation

between the intensity and concentration in D band and 2D band as seen in Figure 3.13 b, meaning

that the deposition of tin oxide does not interact with defects in carbon nanotubes.

Figure 3.13: a. RBM, b. D and 2D band of Raman spectra of carbon nanotube networks deposited

with SnOx with concentrations from 0.25 to 10% w/v TIP. These spectra are included to show that

they have no correlation with dopant concentration.

29

Since the spectra are not from individual nanotubes but from nanotube network, it would

be difficult to study metallic and semiconducting nanotubes separately. However, a quantitative

intensity change is observed in the BWF lineshape of metallic nanotubes, which is shown as the

shoulder located at about 1550 cm-1 as seen in Figure 3.14. The G band peaks of semiconducting

nanotubes are known to be located at ~1580 cm-1,45 the peak at around 1560 cm-1 is also analyzed.

Figure 3.14 b shows the change in position of the peak at 1560 cm-1 and the normalized

intensity of the BWF peaks as functions of TIP concentration used. The peak around 1560 cm-1

blue-shifts by 2 cm-1 as 0.25% w/v TIP is deposited. This peak further blue-shifts until the

concentration reaches 1% w/v, and is stationary at that position for higher TIP concentration used.

The height of the shoulder, indicated by normalized intensity in this figure, decreases down by 25%

as the concentration increases up to 2% w/v. For higher TIP concentration the height increases by

about 10%. This implies that for the case of m-SWNTs, charge transfer doping is maximized with

deposition of 1 to 2% w/v of TIP, which coincides with what was previously demonstrated from

electrical measurement and near-IR absorption spectra.

Figure 3.14: a. G band spectra of nanotube network with different concentration of SnOx deposited.

b. The position of the peak at around 1560 cm-1 (open black square) and the intensity of the

shoulder at around 1550 cm-1 (open blue circle) are plotted as functions of TIP concentration used.

30

In concert, from DC/OP ratio calculation, near-IR absorption and Raman spectra, it is

learned that the TIP concentration between 0.5 and 2% w/v maximizes charge transfer doping and

sheet resistance reduction. Further increase in concentration leads to slight deterioration in

conductance and charge transfer doping. This is thought to be because electronic properties of tin

(IV) oxide nanoparticles are dependent upon particle size. For example, the size-dependent energy

band gap change of sol-gel synthesized tin dioxide particles was studied by Gu and colleagues in

some extent.61 This doping mechanism exploits the misalignment of work functions between

carbon nanotubes and metal oxide nanoparticles. Then, these results indicate that the smaller the

nanoparticles are or the thinner the metal oxide film on the nanotube sidewall, the higher the work

function of metal oxide is. Particle sizes with respect to the metal alkoxide concentration was not

measured in this study. However, further study has to be conducted in order to prove this

hypothesis.

One of the motivations of the nanotube network and metal oxide hybrid system was the

stability. Functionalization of carbon nanotubes or adsorption of small molecules on nanotubes

makes the nanotube films susceptible to extreme environments, whereas it is commonly known

that oxides are far more stable. For stability tests, control (CNTs only), nitric acid exposed (HNO3

doped), tin isopropoxide coated (TIP doped) and nitric acid exposed and tin isopropoxide

passivated (HNO3 & TIP) samples were stored in 90% relative humidity (RH) or in a 130°C oven

for 1 hour, and their sheet resistances were compared before and after the test as shown in Table

3.5 and Table 3.6. As a normalization, sheet resistance change is presented as the ratio between

after and before the test. Carbon nanotube network exposed to nitric acid was passivated with tin

isopropoxide spin-coating, because it is learned that nitric acid treatment largely decreases sheet

resistance whereas the stability under high humidity and heat is known to be inferior.

31

Control samples and TIP doped CNT films show similar magnitude of sheet resistance

increase for both of the tests, where, as expected, the resistance of nitric acid treated films doubled

after 1 hour in 90% RH and tripled after 1 hour in a 130°C oven. Passivation of nitric acid treated

film with TIP reduced the sheet resistance rise; however, the increase is still larger than that of

control samples and TIP doped films.

These test results, along with optical transmittance, were converted into DC/OP ratio for

pristine (undoped), nitric acid treated and tin (IV) oxide deposited carbon nanotube films for better

comparison. Figure 3.15 shows the DC/OP ratio of nanotube films before and after each test. First,

before performing stability tests, indicated by ‘As-prepared’ in the figure, the DC/OP ratios of

nitric acid treated and SnOx deposited films are comparable. The resulting DC/OP ratio for both

sets of samples is between 4 and 5, and this suggests that SnOx deposited nanotube films is

susceptible to a humid environment. Three sets of as-prepared nanotube films mentioned above

Table 3.5: High humidity test results presented in sheet resistance change before and after stored

in 90% relative humidity for 1 hour.

Table 3.6: Heat tolerance test result also in sheet resistances and their ratio before and after stored

in a 130°C oven for 1 hour.

32

were stored in a 130°C oven, and their DC/OP ratios were derived. While the ratio of nitric acid

treated films decreased by a factor of three to four, that of SnOx deposited films was reduced only

about 20%, suggesting higher stability of metal oxide deposited films as expected.

In this chapter, optical and electrical measurements and spectroscopic study have revealed

that TIP concentration of 0.5 to 2% w/v leads to maximum doping level change in sheet resistance

reduction in SWNT films while maintaining optical transmittance. Further, SnOx deposited

nanotube film is far more stable under high temperature than nitric acid treated films, which has

been widely used doping method for carbon nanotubes and graphene as discussed in the

Introduction. This SnOx deposition was applied not only on carbon nanotube networks, but also

on large-area SLG grown by CVD. The stability of nitric acid exposed SLG and TIP doped SLG

are compared along with control, undoped, SLG film as discussed in the next chapter.

As-prepared Stored in

90% RH

for 1h

(Humidity test)

Stored in

130C oven

for 1h

(Heating test)

2

4

6

8

10

Undoped

SnOx coated

HNO3 treated

DC

/OP

ratio (

a.u

.)

Figure 3.15: DC/OP ratio of undoped (black square), SnOx

deposited (red circle) and nitric acid treated (blue triangle)

carbon nanotube films before and after high humidity and high

temperature stability tests.

33

CHAPTER 4

SINGLE-LAYER GRAPHENE TRANSPARENT CONDUCTING FILM

In this study, single-layer graphene films were grown on Cu film via chemical vapor

deposition, and transferred onto glass slides to analyze optical transmittance as well as sheet

resistance before and after deposition of SnOx. However, the widely known optical transmittance

of graphene is about 97.5%, and spin-coating TIP onto SLG did not make any measurable change

in transmittance. This is thought to be because the surface of SLG is far smoother than that of

carbon nanotube film, so the amount of SnOx in unit area left on SLG after spin-coating TIP is far

less than the case of nanotube network.

Whereas optical transmittance was virtually unchanged, SnOx deposition reduced the sheet

resistance of SLG roughly by a factor of two. Figure 4.1 shows the decrease in sheet resistance of

SLG as a function of TIP concentration used. Sheet resistance decreased by 40% as soon as only

0.25% w/v TIP was spin-coated. It further decreased by 48% when the concentration reaches 1%

0.0 0.5 1.0 1.5 2.0

0

10

20

30

40

50

Pe

rce

nt

De

cre

ase

in

Sh

ee

t R

esis

tan

ce

(%

)

Tin Isopropoxide Concentration (% w/v)

Figure 4.1: Decrease in sheet resistance of CVD SLG versus TIP

concentration.

34

w/v, and the decrease is practically saturated until 2% w/v TIP. This trend is precisely what was

observed previously in carbon nanotube network and tin (IV) oxide system.

Raman spectra of single-layer graphene before and after TIP deposition are shown in

Figure 4.2. The G band peak position shift and the change in 2D/G intensity ratio are of our interest,

because they indicate change in electronic properties in the system. For example, Basko and

colleagues studied the relationship between the doping level and the intensity in 2D band in

graphene,42 and Nguyen and colleagues analyzed gate-voltage-dependent G band peak position in

graphene.8 The G band was shifted by 9 cm-1, and the 2D/G intensity ratio became roughly half,

meaning that SLG is strongly doped.

Further, the stability in resistance against heat was tested as shown in Figure 4.3. Three

sets of samples, control, nitric acid treated and tin isopropoxide coated samples, were heated with

an 80°C hot plate for 2 hours. The resistance of all of the films increased by practically the same

magnitude. Careful surface engineering is required to ensure the stability under high humidity and

Figure 4.2: Raman spectra of SLG before and after spin-coating

0.5% w/v TIP.

35

heat, because, unlike carbon nanotube network, the surface of SLG would be 2-dimensional plane.

Exposing graphene to environment might lead to instability in its electronic properties, where fully

passivating the surface would introduce additional resistance between graphene and contact or

other layers.

For application purposes, it is beneficial that transmittance does not change and resistance

is reduced after spin-coating metal alkoxide solution. However, electronic properties of this system

also have to be analyzed, which can be done by ultraviolet photoelectron spectroscopy (UPS).

Carbon nanotube network and tin (IV) oxide hybrid system was optically, electrically and

spectroscopically studied, and for potential application purposes, a proof-of-concept carbon-

nanotube-based thin film diode was fabricated.

0.0 0.5 1.0 1.5 2.0

1.0

1.1

1.2

1.3

1.4

1.5

1.6

Rs(a

fter)

/ R

s(b

efo

re)

(a.u

.)

Heated on 80C hot plate (hours)

Undoped

Tin isopropoxide spin-coated

Exposure to nitric acid vapor

Figure 4.3: Change in sheet resistance in ratio as a function of

time on an 80°C hot plate.

36

CHAPTER 5

POTENTIAL APPLICATION

Carbon-based electrodes have been extensively studied in optoelectronic application.

Carbon nanotube network has been used as transparent electrodes for organic solar cells62 or light-

emitting diodes.63,64 However, it is widely learned that carbon-based electrodes have higher sheet

resistance by orders of magnitude compared to conventionally used ITO, as discussed in

Introduction. One of the purposes of this research was to reduce inherently high resistance of

carbon nanotube network by charge transfer doping. In order to test the possibility of carbon

nanotube and tin (IV) oxide hybrid electrode, a proof-of-concept thin film diode based on was

fabricated.

A schematic diagram of this device is shown in Figure 5.1 a, and the corresponding band

diagram is presented in Figure 5.1 b. Tin isopropoxide solution was spin-coated on carbon

nanotube film, and additional tin dioxide powder was deposited as the hole blocking layer. P3OT

film was used as the electron blocking layer, and gold, used as cathode in this device, was

evaporated in a separate substrate to prevent current leakage. Figure 5.1 c and d are current density

versus voltage curves in linear and log scales, respectively. From the logarithmic lineshape in the

linear plot and the linear trend in the log plot, it is concluded that a thin film diode based on carbon

nanotube film has been successfully constructed.

Even though both plots show that the device exhibits rectifying behavior of diodes, there

are issues in the device: current is too small, current level increase with light is only about a factor

of three and the turn-on voltage is huge. The primary problem is thought to be that active region

in the diode is tiny. This hypothesis is drawn from the fact that the device was not functioning

when gold cathode was directly evaporated onto the device, and that tin dioxide powder layer was

37

not completely covering the carbon nanotube electrode. When observed with a scanning electron

microscope, as shown in Figure 5.2, tin dioxide particles are not entirely covering the nanotube

network although the thickness is on the order of micrometer. A well-dispersed sol-gel synthesized

metal oxide nanoparticles can be spin-coated instead; however, the thickness is known to be

roughly 20 nm where the roughness of nanotube network is larger than that. Therefore, conformal

coating of powder on nanotube film should be achieved not only to prevent electrical shortage but

also to improve device performance.

Based on this carbon nanotube based diode, a photovoltaic device was also constructed by

spin-coating or drop-drying CdSe/CdTe nanorod heterostructure (NRH) solution between SnO2

Figure 5.1: a. A schematic diagram of carbon nanotube-based thin film diode. b. A band diagram

of this thin film diode. c. Dark current (open black square) and photocurrent (open red circle) per

device area versus voltage curves (JV curve) of thin film diode. d. The JV curve in c is re-plotted

in log scale to show the diode behavior.

38

and P3OT layers. Here, fluorine-doped tin oxide (FTO) electrode and TiO2 electron transport layer

(ETL) of the device constructed by McDaniel and colleagues are substituted by carbon nanotube

network electrode and SnO2 ETL.53 A schematic band diagram is shown in Figure 5.3 a, and the

dark JV curve with this solar cell is presented in Figure 5.3 b. The magnitude of current density

and applied voltage is similar to the thin film diode without the NRH active layer. Not only was

the current level small and the necessary applied voltage large, but also the operation was unstable.

Most of the devices only occasionally exhibited rectifying behavior, and the same device would

often show a linear JV curve with high current level, meaning the device was functioning or

shorted at different measurements.

Photocurrent was also measured with ambient light as presented in Figure 5.4, and the JV

curve is magnified around zero current density and zero applied voltage in an attempt to observe

open-circuit voltage and short-circuit current density. However, both were effectively zero, and

this would mean that charge separation does not take place in this device.

It is believed that the primary problem for unstable operation is the tin (IV) dioxide particle

layer not entirely covering the carbon nanotube electrode and subsequent layer, as it was for some

Figure 5.2: a. A birds-eye view of carbon nanotube film with SnO2 film. The bright side is the part

covered with SnO2 nanoparticles. b. A cross-sectional view of carbon nanotube film and the layer

of SnO2 particles. Note that the figure is upside down.

39

of the thin-film diode samples. Suppose that the ETL is partially covering the carbon nanotube

network and the HTL and the cathode are constructed only on top of that area. Then, even though

the device shows a rectifying behavior, nominal active device area would be too small, leading to

low current level per projected area and high operation voltage. On the other hand, if the ETL

partially covers the nanotube electrode and either the HTL or the cathode makes contact to

nanotube network, then the device would be shorted.

Figure 5.3: a. The expected band diagram of the photovoltaic device with nanorod heterojunction

active layer and carbon nanotube electrode. b. The dark JV curve obtained from this solar cell.

Figure 5.4: a. The JV curve of the solar cell with ambient light. b. This JV curve is magnified

around zero current density and zero applied voltage.

40

As explained in Experimental Methods, devices were constructed from carbon nanotube

electrode up to HTL on one glass substrate, the Au cathode was separately evaporated on the other

glass substrate and they were put together with a binder clip. Depending on how strongly the

substrates are bound together, the ‘active area’ of the device would be changing, leading to the

problems described above. Therefore, same as the previous thin-film diode case, it must be ensured

that each layer is entirely covering the surface. Direct observation of the surface by SEM or optical

microscopy would be good methods to investigate the coverage.

41

CHAPTER 6

CONCLUSION

In conclusion, carbon nanotube network and SnOx hybrid systems were studied. First,

various anionic, cationic and nonionic surfactants were used to disperse SWNTs in deionized water

to choose the right surfactant for transparent conducting film applications, and different metal

alkoxides were applied on nanotube films to choose the best metal oxide for dopant. Optical

transmittance and electrical measurement data were used to calculate one figure of merit for each

case, DC/OP ratio, and finally SDBS and tin isopropoxide was selected to be the best dispersant

and dopant metal alkoxide, respectively. CVD growth of carbon nanotubes were attempted in order

to obtain low-resistance undoped carbon nanotube films; however, both Fe/Mo/Al2O3 CCVD and

FCCVD were not successful compared to arc-discharge CNTs.

Carbon nanotube film and tin (IV) oxide system was studied with optical and electrical

measurements and near-IR absorption and Raman spectroscopies with respect to TIP concentration,

and the concentration of 0.5 to 2% w/v was shown to be minimizing the film resistance and

maximizing charge transfer doping. This TIP deposition was applied on single-layer graphene, and

the resistance was decreased as low as a factor of two, while transmittance is practically unaffected.

The stability of this system against high temperature at around 130°C was substantially better than

its competitive doping method, nitric acid treatment. The performance, represented as DC/OP ratio,

was decreased roughly 20% after the stability test, whereas the one of nitric acid treated nanotube

film plunged by a factor of four.

A proof-of-concept thin film diode was fabricated with nanotube film and SnOx system,

which did show a rectifying behavior. The nanotube/SnOx system was also applied to a CdSe/CdTe

NRH solar cell as the electrode, which occasionally exhibited rectifying behavior but low current

42

and high operation voltage. Partial coverage of the SnO2 nanoparticle ETL is hypothetically the

cause of the unstable operation as qualitatively discussed above. Therefore, conformal coating of

carbon nanotube network has to be ensured for transparent electrode applications. If conformal

coating can be achieved, this system can be widely applied in not only solar cells but also light-

emitting diodes. We have methods to quantify the performance of our carbon-based thin film as

discussed previously, the percolation in carbon nanotube network.48 Further, researchers are

developing manufacturing methods for continuous production of carbon nanotube or graphene

films.19,34,65 Therefore, it is fair to state that carbon-based electrodes have a bright future ahead for

research, practical application and mass production.

43

REFERENCES

(1) Iijima, S. Helical Microtubules of Graphitic Carbon. Nature 1991, 354, 56–58.

(2) Novoselov, K. S.; Geim, a K.; Morozov, S. V; Jiang, D.; Katsnelson, M. I.; Grigorieva, I.

V; Dubonos, S. V; Firsov, a a. Two-Dimensional Gas of Massless Dirac Fermions in

Graphene. Nature 2005, 438, 197–200.

(3) Edwards, P. P.; Porch, a; Jones, M. O.; Morgan, D. V; Perks, R. M. Basic Materials

Physics of Transparent Conducting Oxides. Dalton Trans. 2004, 2995–3002.

(4) Ginley, D. S.; Bright, C. Transparent Conducting Oxides. MRS Bull. 2011, 25, 15–18.

(5) Caruge, J. M.; Halpert, J. E.; Wood, V.; Bulović, V.; Bawendi, M. G. Colloidal Quantum-

Dot Light-Emitting Diodes with Metal-Oxide Charge Transport Layers. Nat. Photonics

2008, 2, 247–250.

(6) Oh, N.; Nam, S.; Zhai, Y.; Deshpande, K.; Trefonas, P.; Shim, M. Double-Heterojunction

Nanorods. Nat. Commun. 2014, 5, 3642.

(7) Cui, J.; Wang, A.; Edleman, N. L.; Ni, J.; Lee, P.; Armstrong, N. R.; Marks, T. J. Indium

Tin Oxide Alternatives-High Work Function Transparent Conducting Oxides as Anodes

for Organic Light-Emitting Diodes. Adv. Mater. 2001, 13, 1476–1480.

(8) Nguyen, K. T.; Abdula, D.; Tsai, C.-L.; Shim, M. Temperature and Gate Voltage

Dependent Raman Spectra of Single-Layer Graphene. ACS Nano 2011, 5, 5273–5279.

(9) Nguyen, K.; Gaur, A.; Shim, M. Fano Lineshape and Phonon Softening in Single Isolated

Metallic Carbon Nanotubes. Phys. Rev. Lett. 2007, 98, 145504.

(10) Back, J.; Tsai, C.-L.; Kim, S.; Mohammadi, S.; Shim, M. Manifestation of Kohn Anomaly

in 1/f Fluctuations in Metallic Carbon Nanotubes. Phys. Rev. Lett. 2009, 103, 215501.

(11) Shin, D.-W.; Lee, J. H.; Kim, Y.-H.; Yu, S. M.; Park, S.-Y.; Yoo, J.-B. A Role of HNO3

on Transparent Conducting Film with Single-Walled Carbon Nanotubes. Nanotechnology

2009, 20, 475703.

(12) Abdula, D.; Shim, M. Performance and Photovoltaic Response of Polymer-Doped Carbon

Nanotube P-N Diodes. ACS Nano 2008, 2, 2154–2159.

(13) Gaur, A.; Shim, M. Substrate-Enhanced O2 Adsorption and Complexity in the Raman G-

Band Spectra of Individual Metallic Carbon Nanotubes. Phys. Rev. B 2008, 78, 125422.

44

(14) Abdula, D.; Ozel, T.; Kang, K.; Cahill, D. G.; Shim, M. Environment-Induced Effects on

the Temperature Dependence of Raman Spectra of Single-Layer Graphene. J. Phys.

Chem. C 2008, 112, 20131–20134.

(15) Wu, Z.; Chen, Z.; Du, X.; Logan, J. M.; Sippel, J.; Nikolou, M.; Kamaras, K.; Reynolds, J.

R.; Tanner, D. B.; Hebard, A. F.; et al. Transparent, Conductive Carbon Nanotube Films.

Science 2004, 305, 1273–1276.

(16) Song, Y. Il; Yang, C.-M.; Kim, D. Y.; Kanoh, H.; Kaneko, K. Flexible Transparent

Conducting Single-Wall Carbon Nanotube Film with Network Bridging Method. J.

Colloid Interface Sci. 2008, 318, 365–371.

(17) Sun, J.; Wang, R. Carbon Nanotube Transparent Electrode. 2013.

(18) Mora, A.; Khan, K. A.; El Sayed, T. Crack Density and Electrical Resistance in Indium-

Tin-Oxide/polymer Thin Films under Cyclic Loading. Electron. Mater. Lett. 2014, 10,

1033–1037.

(19) Bae, S.; Kim, H.; Lee, Y.; Xu, X.; Park, J.-S.; Zheng, Y.; Balakrishnan, J.; Lei, T.; Kim,

H. R.; Song, Y. Il; et al. Roll-to-Roll Production of 30-Inch Graphene Films for

Transparent Electrodes. Nat. Nanotechnol. 2010, 5, 574–578.

(20) Bekyarova, E.; Itkis, M. E.; Cabrera, N.; Zhao, B.; Yu, A.; Gao, J.; Haddon, R. C.

Electronic Properties of Single-Walled Carbon Nanotube Networks. J. Am. Chem. Soc.

2005, 127, 5990–5995.

(21) Kim, H.; Pique, A.; Horwitz, J. S.; Mattoussi, H.; Murata, H.; Kafafi, Z. H.; Chrisey, D. B.

Indium Tin Oxide Thin Films for Organic Light-Emitting Devices. Appl. Phys. Lett. 1999,

74, 3444.

(22) Gao, J.; Wang, W. Y.; Cui, L. J.; Chen, L. T.; Hu, X. Y.; Li, H.; Geng, H. Z. Effect of

Different Concentrations of Nitric Acid on the Conductivity of Single-Walled Carbon

Nanotube Transparent Films. Adv. Mater. Res. 2013, 658, 3–7.

(23) Kamarás, K.; Pekker, Á.; Botka, B.; Hu, H.; Niyogi, S.; Itkis, M. E.; Haddon, R. C. The

Effect of Nitric Acid Doping on the Optical Properties of Carbon Nanotube Films. Phys.

Status Solidi 2010, 247, 2754–2757.

(24) Parekh, B. B.; Fanchini, G.; Eda, G.; Chhowalla, M. Improved Conductivity of

Transparent Single-Wall Carbon Nanotube Thin Films via Stable Postdeposition

Functionalization. Appl. Phys. Lett. 2007, 90, 121913.

(25) Kim, S. M.; Kim, K. K.; Jo, Y. W.; Park, M. H.; Chae, S. J.; Duong, D. L.; Yang, C. W.;

Kong, J.; Lee, Y. H. Role of Anions in the AuCl3-Doping of Carbon Nanotubes. ACS

Nano 2011, 5, 1236–1242.

45

(26) Kim, K. K.; Bae, J. J.; Park, H. K.; Kim, S. M.; Geng, H.-Z.; Park, K. A.; Shin, H.-J.;

Yoon, S.-M.; Benayad, A.; Choi, J.-Y.; et al. Fermi Level Engineering of Single-Walled

Carbon Nanotubes by AuCl3 Doping. J. Am. Chem. Soc. 2008, 130, 12757–12761.

(27) Kim, K. K.; Reina, A.; Shi, Y.; Park, H.; Li, L.-J.; Lee, Y. H.; Kong, J. Enhancing the

Conductivity of Transparent Graphene Films via Doping. Nanotechnology 2010, 21,

285205.

(28) Lee, H.; Lee, J.; Park, S.; Yi, Y.; Cho, S. W.; Kim, J. W.; Kang, S. J. Hole Injection

Enhancement of a Single-Walled Carbon Nanotube Anode Using an Organic Charge-

Generation Layer. Carbon N. Y. 2014, 71, 268–275.

(29) Huang, J.-H.; Fang, J.-H.; Liu, C.-C.; Chu, C.-W. Effective Work Function Modulation of

Graphene/carbon Nanotube Composite Films as Transparent Cathodes for Organic

Optoelectronics. ACS Nano 2011, 5, 6262–6271.

(30) Castro Neto, a. H.; Peres, N. M. R.; Novoselov, K. S.; Geim, a. K. The Electronic

Properties of Graphene. Rev. Mod. Phys. 2009, 81, 109–162.

(31) Mayorov, A. S.; Elias, D. C.; Mukhin, I. S.; Morozov, S. V; Ponomarenko, L. A.;

Novoselov, K. S.; Geim, A. K.; Gorbachev, R. V. How Close Can One Approach the

Dirac Point in Graphene. 2012.

(32) Geim, A. K.; Novoselov, K. S. The Rise of Graphene. Nat. Mater. 2007, 6, 183–191.

(33) Malard, L. M.; Pimenta, M. a.; Dresselhaus, G.; Dresselhaus, M. S. Raman Spectroscopy

in Graphene. Phys. Rep. 2009, 473, 51–87.

(34) Ryu, J.; Kim, Y.; Won, D.; Kim, N.; Park, J. S.; Lee, E.-K.; Cho, D.; Cho, S.-P.; Kim, S.

J.; Ryu, G. H.; et al. Fast Synthesis of High-Performance Graphene Films by Hydrogen-

Free Rapid Thermal Chemical Vapor Deposition. ACS Nano 2014, 8, 950–956.

(35) Dresselhaus, M. S.; Dresselhaus, G.; Saito, R.; Jorio, a. Raman Spectroscopy of Carbon

Nanotubes. Phys. Rep. 2005, 409, 47–99.

(36) Lin, Y.-M.; Valdes-Garcia, A.; Han, S.-J.; Farmer, D. B.; Meric, I.; Sun, Y.; Wu, Y.;

Dimitrakopoulos, C.; Grill, A.; Avouris, P.; et al. Wafer-Scale Graphene Integrated

Circuit. Science 2011, 332, 1294–1297.

(37) Avouris, P.; Freitag, M.; Perebeinos, V. Carbon-Nanotube Photonics and Optoelectronics.

Nat. Photonics 2008, 2, 341–350.

(38) Sgobba, V.; Guldi, D. M. Carbon Nanotubes--Electronic/electrochemical Properties and

Application for Nanoelectronics and Photonics. Chem. Soc. Rev. 2009, 38, 165–184.

46

(39) Kataura, H.; Kumazawa, Y.; Maniwa, Y.; Umezu, I.; Suzuki, S.; Ohtsuka, Y.; Achiba, Y.

Optical Properties of Single-Wall Carbon Nanotubes. Synth. Met. 1999, 103, 2555–2558.

(40) Sinnott, S. B.; Andrews, R. Carbon Nanotubes: Synthesis, Properties, and Applications.

Crit. Rev. Solid State Mater. Sci. 2001, 26, 145–249.

(41) O’ Connell, M. J.; Eibergen, E. E.; Doorn, S. K. Chiral Selectivity in the Charge-Transfer

Bleaching of Single-Walled Carbon-Nanotube Spectra. Nat. Mater. 2005, 4, 412–418.

(42) Basko, D. M.; Piscanec, S.; Ferrari, a. C. Electron-Electron Interactions and Doping

Dependence of the Two-Phonon Raman Intensity in Graphene. Phys. Rev. B 2009, 80,

165413.

(43) Ferrari, A. C.; Basko, D. M. Raman Spectroscopy as a Versatile Tool for Studying the

Properties of Graphene. Nat. Nanotechnol. 2013, 8, 235–246.

(44) Eklund, P. C.; Holden, J. M.; Jishi, R. A. Vibrational Modes of Carbon Nanotubes;

Spectroscopy and Theory. Carbon N. Y. 1995, 33, 959–972.

(45) Dresselhaus, M. S.; Jorio, A.; Hofmann, M.; Dresselhaus, G.; Saito, R. Perspectives on

Carbon Nanotubes and Graphene Raman Spectroscopy. Nano Lett. 2010, 10, 751–758.

(46) Shim, M.; Gaur, A.; Nguyen, K. T.; Abdula, D.; Ozel, T. Spectral Diversity in Raman G-

Band Modes of Metallic Carbon Nanotubes within a Single Chirality. J. Phys. Chem. C

2008, 112, 13017–13023.

(47) Shim, M.; Gaur, A.; Nguyen, K. T.; Abdula, D.; Ozel, T. Spectral Diversity in Raman G-

Band Modes of Metallic Carbon Nanotubes within a Single Chirality. J. Phys. Chem. C

2008, 112, 13017–13023.

(48) Hu, L.; Hecht, D.; Grüner, G. Percolation in Transparent and Conducting Carbon

Nanotube Networks. Nano Lett. 2004, 0–4.

(49) Hecht, D. S.; Hu, L.; Irvin, G. Emerging Transparent Electrodes Based on Thin Films of

Carbon Nanotubes, Graphene, and Metallic Nanostructures. Adv. Mater. 2011, 23, 1482–

1513.

(50) Baeumer, C.; Rogers, S. P.; Xu, R.; Martin, L. W.; Shim, M. Tunable Carrier Type and

Density in graphene/PbZr0.2Ti0.8O3 Hybrid Structures through Ferroelectric Switching.

Nano Lett. 2013, 13, 1693–1698.

(51) Cassell, A. M.; Raymakers, J. A.; Kong, J.; Dai, H. Large Scale CVD Synthesis of Single-

Walled Carbon Nanotubes. J. Phys. Chem. B 1999, 103, 6484–6492.

47

(52) Liu, Q.; Ren, W.; Chen, Z.-G.; Wang, D.-W.; Liu, B.; Yu, B.; Li, F.; Cong, H.; Cheng, H.-

M. Diameter-Selective Growth of Single-Walled Carbon Nanotubes with High Quality by

Floating Catalyst Method. ACS Nano 2008, 2, 1722–1728.

(53) McDaniel, H.; Heil, P. E.; Tsai, C.-L.; Kim, K. K.; Shim, M. Integration of Type II

Nanorod Heterostructures into Photovoltaics. ACS Nano 2011, 5, 7677–7683.

(54) Zhang, C.; Valsaraj, K. T.; Constant, W. D.; Roy, D. Aerobic Biodegradation Kinetics of

Four Anionic and Nonionic Surfactants at Sub- and Supra-Critical Micelle Concentrations

(CMCs). Water Res. 1999, 33, 115–124.

(55) Minami, T. New N-Type Transparent Conducting Oxides. MRS Bull. 2011, 25, 38–44.

(56) Kröger, M.; Hamwi, S.; Meyer, J.; Riedl, T.; Kowalsky, W.; Kahn, A. P-Type Doping of

Organic Wide Band Gap Materials by Transition Metal Oxides: A Case-Study on

Molybdenum Trioxide. Org. Electron. 2009, 10, 932–938.

(57) Agekyan, V. T. Spectroscopic Properties of Semiconductor Crystals with Direct

Forbidden Energy Gap. Phys. Status Solidi 1977, 43, 11–42.

(58) Stouwdam, J. W.; Janssen, R. A. J. Red, Green, and Blue Quantum Dot LEDs with

Solution Processable ZnO Nanocrystal Electron Injection Layers. J. Mater. Chem. 2008,

18, 1889.

(59) Irfan, I.; James Turinske, A.; Bao, Z.; Gao, Y. Work Function Recovery of Air Exposed

Molybdenum Oxide Thin Films. Appl. Phys. Lett. 2012, 101, 093305.

(60) Kwak, J.; Bae, W. K.; Lee, D.; Park, I.; Lim, J.; Park, M.; Cho, H.; Woo, H.; Yoon, D. Y.;

Char, K.; et al. Bright and Efficient Full-Color Colloidal Quantum Dot Light-Emitting

Diodes Using an Inverted Device Structure. Nano Lett. 2012, 12, 2362–2366.

(61) Gu, F.; Wang, S. F.; Lü, M. K.; Zhou, G. J.; Xu, D.; Yuan, D. R. Photoluminescence

Properties of SnO 2 Nanoparticles Synthesized by Sol−Gel Method. J. Phys. Chem. B

2004, 108, 8119–8123.

(62) Rowell, M. W.; Topinka, M. a.; McGehee, M. D.; Prall, H.-J.; Dennler, G.; Sariciftci, N.

S.; Hu, L.; Gruner, G. Organic Solar Cells with Carbon Nanotube Network Electrodes.

Appl. Phys. Lett. 2006, 88, 233506.

(63) Li, J.; Hu, L.; Wang, L.; Zhou, Y.; Grüner, G.; Marks, T. J. Organic Light-Emitting

Diodes Having Carbon Nanotube Anodes. Nano Lett. 2006, 6, 2472–2477.

(64) Zhang, D.; Ryu, K.; Liu, X.; Polikarpov, E.; Ly, J.; Tompson, M. E.; Zhou, C.

Transparent, Conductive, and Flexible Carbon Nanotube Films and Their Application in

Organic Light-Emitting Diodes. Nano Lett. 2006, 6, 1880–1886.

48

(65) Dan, B.; Irvin, G. C.; Pasquali, M. Continuous and Scalable Fabrication of Transparent

Conducting Carbon Nanotube Films. ACS Nano 2009, 3, 835–843.