Solution-deposited organic - Inorganic hybrid multilayer gate dielectrics. Design, synthesis, microstructures, and electrical properties with thin-film transistors

Article abstract of DOI:10.1021/ja202755x

We report here on the rational synthesis, processing, and dielectric properties of novel layer-by-layer organic/inorganic hybrid multilayer dielectric films enabled by polarizable π-electron phosphonic acid building blocks and ultrathin ZrO₂ layers. These new zirconia-based self-assembled nanodielectric (Zr-SAND) films (5 - 12 nm thick) are readily fabricated via solution processes under ambient atmosphere. Attractive Zr-SAND properties include amenability to accurate control of film thickness, large-area uniformity, well-defined nanostructure, exceptionally large electrical capacitance (up to 750 nF/cm²), excellent in ulating properties (leakage current densities as low as 10^{- 7} A/cm²), and excellent thermal stability. Thin-film transistors (TFTs) fabricated with pentacene and PDIF-CN₂ as representative organic semiconductors and zinc - tin - oxide (Zn - Sn - O) as a representative inorganic semiconductor function well at low voltages (<±4.0 V). Furthermore, the TFT performance parameters of representative organic semiconductors deposited on Zr-SAND films, functionalized on the surface with various alkylphosphonic acid self-assembled monolayers, are investigated and shown to correlate closely with the alkylphosphonic acid chain dimensions.

Full text of DOI:10.1021/ja202755x

ARTICLE
pubs.acs.org/JACS
Solution-Deposited OrganicꢀInorganic Hybrid Multilayer Gate
Dielectrics. Design, Synthesis, Microstructures, and Electrical
Properties with Thin-Film Transistors
||
Young-geun Ha,† Jonathan D. Emery,‡ Michael J. Bedzyk,‡,§ Hakan Usta,|| Antonio Facchetti,*,†, and
Tobin J. Marks*,†,‡
†Department of Chemistry, ‡Department of Materials Science and Engineering, and the Materials Research Center, and §Department of
Physics and Astronomy, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, United States
Polyera Corporation, 8045 Lamon Avenue, Skokie, Illinois 60077, United States
S Supporting Information
b
ABSTRACT: We report here on the rational synthesis, proces-
sing, and dielectric properties of novel layer-by-layer organic/
inorganic hybrid multilayer dielectric films enabled by polariz-
able π-electron phosphonic acid building blocks and ultrathin
ZrO2 layers. These new zirconia-based self-assembled nanodi-
electric (Zr-SAND) films (5ꢀ12 nm thick) are readily fabri-
cated via solution processes under ambient atmosphere.
Attractive Zr-SAND properties include amenability to accurate
control of film thickness, large-area uniformity, well-defined
nanostructure, exceptionally large electrical capacitance (up to 750 nF/cm2), excellent insulating properties (leakage current
densities as low as 10ꢀ7 A/cm2), and excellent thermal stability. Thin-film transistors (TFTs) fabricated with pentacene and PDIF-
CN2 as representative organic semiconductors and zincꢀtinꢀoxide (ZnꢀSnꢀO) as a representative inorganic semiconductor
function well at low voltages (<(4.0 V). Furthermore, the TFT performance parameters of representative organic semiconductors
deposited on Zr-SAND films, functionalized on the surface with various alkylphosphonic acid self-assembled monolayers, are
investigated and shown to correlate closely with the alkylphosphonic acid chain dimensions.
’ INTRODUCTION
Over the past two decades, solution-processable organic, inor-
ganic, and polymeric semiconductors were developed due to
attractions such as printability, the possibility of large-area deposi-
tions, low-cost device fabrication, and compatibility with mech-
anically flexible substrates.15ꢀ23 Despite recent progress, one prin-
cipal limitation of these semiconductors is their relatively low carrier
mobilities, well below those of most silicon-based high-performance
materials.24ꢀ26 As a result, TFTs fabricated from these semicon-
ductors require high operating voltages to attain usable drain current
(IDS). For low power applications such as RF-ID tags, flat panel
displays, and portable electronics, it is mandatory to achieve high
TFT drain currents (IDS) at acceptably low operating voltages.
Without changing device geometry (W and L) and semiconductor
material (μ), an alternative to overcome these mobility limitations is
to increase the gate dielectric capacitance Ci, given by eq 2:
Thin-film transistors (TFTs) fabricated from unconventional
materials and by unconventional methodologies are of interest for
future low-cost electronic applications such as printed RF-ID cards,
flexible displays, and sensors.1ꢀ14 Transistors are the key compo-
nents used for current modulation and switching in all modern
electronic devices (Figure 1). The basic working principle of TFTs
is that the channel source and drain current (IDS) in saturation is
modulated by the source-gate bias (VG) according to eq 1:
W
2
IDS
¼
μCiðVG ꢀ VTÞ
ð1Þ
2L
where W/L is the channel width/length, Ci is the dielectric
capacitance per unit area, μ is the charge carrier mobility, VG is
the source-gate voltage, and VT is the threshold voltage. Depending
on the charge carrier sign in the channel between source and drain,
the semiconductor is either hole- (p-type) or electron-transporting
(n-type). The two most important parameters governing TFT
performance are the field-effect mobility (μ) and the current on/off
ratio (Ion/Ioff). These parameters define the drift velocity of the
charge carriers in the semiconductor layer under the source/drain
electric field and the current modulation between the TFT “on” and
“off” states upon a gate voltage change, respectively.
k
Ci ¼ ε0
ð2Þ
d
From eq 2, note that operating bias reduction can be achieved by
either increasing the dielectric constant (k) or decreasing the
Received: March 27, 2011
Published: May 24, 2011
r
2011 American Chemical Society
10239
dx.doi.org/10.1021/ja202755x J. Am. Chem. Soc. 2011, 133, 10239–10250
|

Journal of the American Chemical Society
ARTICLE
of organic materials.29,30 Several groups have reported very low
operating voltage OTFTs using hybrid materials combining self-
assembled monolayers (SAMs) with ultrathin MO layers such as
those of HfOx, AlOx, and ZrOx.31ꢀ36 However, the SAM
thicknesses of these hybrid films are limited by the singly
functionalized self-assembly precursors, and it has been impos-
sible to fabricate multilayers with well-defined growth character-
istics. A key to utilizing multilayered organicꢀinorganic hybrid
materials in unconventional TFTs and other applications is the
ability to prepare high-quality multilayers in the simplest and
most reliable manner. Vapor-phase fabrication methods for
organicꢀinorganic hybrid materials are promising approaches
to high-quality hybrid films;37ꢀ39 however, they typically require
high- or medium-vacuum deposition equipment (e.g., atomic
layer or chemical/physical vapor deposition), and low-cost path-
ways for integration into large-volume coating processes are not
obvious. Layer-by-layer solution-based deposition of well-
defined organic precursors allows the realization of a range of
functional materials with a high degree of order and structural
control at the molecular level. A variety of such assemblies
have been reported, often based on siloxane chemistry40,41 or
on metalꢀligand coordination.42ꢀ46 Among these materials
systems, this laboratory reported a class of silane-based nanodi-
electrics called self-assembled nanodielectrics (SANDs) consist-
ing of alternating organicꢀinorganic multilayers, chemically
bound by strong siloxane linkages, and offering promising
properties for a variety of opto-electronic applications, including
TFTs.47ꢀ51 However, this self-assembly approach uses ambient-
sensitive and highly reactive chlorosilane reagents, which require
anhydrous atmospheres and manipulation to control their
chemistry. Therefore, the question arises as to whether it might
Figure 1. Schematic of a bottom-gate top-contact thin-film transistor
(TFT) device geometry and the zirconia-phosphonate self-assembled
nanodielectric (Zr-SAND) employed in this study.
thickness (d) of the gate dielectric layer. An attractive approach is to
employ high-k materials such as metal oxide (MO) films; however,
high-quality MO dielectric films typically require high growth/
annealing temperatures (>400 °C) and/or expensive vacuum
deposition technologies to ensure acceptably low leakage
currents.27,28 Therefore, a promising and challenging approach to
lower the operating voltages is to use solution-processable high-k
dielectrics fabricated at low temperature, which afford high capaci-
tance values as well as low leakage currents.
Organicꢀinorganic hybrid thin films are of interest for
advanced electronics applications due to the potential for com-
bining the distinctive properties of both the organic and the
inorganic components. Hybrid materials provide both the op-
tical, electrical, and environmental durability of inorganic materi-
als, as well as the mechanical flexibility and properties tunability
Scheme 1. Synthesis of Polarizable Phosphonic Acid-Based π-Electron Building Block
10240
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
be possible to use alternative self-assembly strategies to fabricate
high-performance multilayer gate dielectrics for TFT applica-
tions. In this contribution, we demonstrate a new approach to
organicꢀinorganic hybrid multilayer gate dielectrics for OTFTs
using polarizable, phosphonic acid-functionalized organic
precursors combined with ultrathin layers of high-k inorganic
oxide materials. It will be seen that these reagents are well-
suited for ambient atmosphere fabrication of self-assembled
multilayers, exhibiting high capacitance (465ꢀ750 nF/cm2)
and low leakage current density (∼10ꢀ7A/cm2 at 2 MV/cm),
grown in a simple and highly reliable manner, providing
uniformly structured multilayers via a solution process.
Furthermore, it will be shown that these dielectrics function
effectively with both representative organic and inorganic
semiconductors and can be modified by various self-as-
sembled monolayers (SAMs) to enhance the TFT perfor-
mance even further.
1.00 mmol) and 4-(bromomethyl)phenylphosphonate (3) (0.317 g,
1.00 mmol) was dissolved in CH2Cl2 (10 mL). The red mixture was
heated under a nitrogen atmosphere at 60 °C for 4 h. The solvent was
then removed under high vacuum, and the residue was dried under high
1
vacuum overnight. H NMR (500 MHz, DMSO-d6): δ 9.07 (d, 2H,
JHꢀH = 6 Hz), 8.14 (d, 2H, JHꢀH = 7 Hz), 7.90 (d, 2H, JHꢀH = 9.5 Hz),
7.76ꢀ7.80 (m, 2H), 7.62ꢀ7.64 (m, 2H), 7.07 (d, 2H, JHꢀH = 9.5 Hz),
5.88 (s, 2H), 4.95ꢀ4.97 (m, 2H), 3.99ꢀ4.03 (m, 4H), 3.99ꢀ4.03 (m,
4H), 3.66ꢀ3.72 (m, 8H), 1.22 (t, 3H). 31P NMR (400 MHz, DMSO-
d6): δ ꢀ18.07 (s, 1P). High resolution EI-MS calcd for C26H34N4O5Pþ:
513.22744. Found: 513.22668.
Synthesis of 4-[[4-[Bis(2-hydroxyethyl)amino]phenyl]diazenyl]-1-
[4-(diethoxyphosphoryl) benzyl]pyridinium Bromide (5). To a solu-
tion of 4-[[4-[bis(2-hydroxyethyl)amino]phenyl]diazenyl]-1-(4-phos-
phonobenzyl) pyridinium bromide (4) (0.297 g, 0.5 mmol) in
anhydrous CH2Cl2 (30 mL) was added trimethylbromosilane (10 equiv,
0.66 mL) dropwise over 10 min. The mixture was then stirred over-
night at room temperature under a nitrogen atmosphere. After
completion of the reaction, the solvent was evaporated and the residue
was dissolved in methanol (5 mL). Filtration and evaporation of the solvent
afforded 0.29 g of the pure final product PAE as a red powder. 1H NMR
’ EXPERIMENTAL SECTION
(500 MHz, DMSO-d6): δ 9.05 (d, 2H, JHꢀH = 7 Hz), 8.12 (d, 2H, JHꢀH
=
Materials. Phosphonic acid-based π-electron (PAE) precursor 5
was straightforwardly synthesized as shown in Scheme 1. Chemical
reagents for synthesis were purchased from Aldrich Chemical Co. unless
otherwise indicated. The precursors for the ZrO2 layers were prepared
using zirconium(IV) chloride (99.5%, Aldrich). Heavily n-doped Si
wafers (Montco Silicon Tech, Inc.) for gate electrodes were cleaned
according to standard procedures. For phosphonic acid-based self-
assembled monolayers (PA-SAMs), n-hexyl-phosphonic acid (PA-C6),
n-dodecylphosphonic acid (PA-C12), n-tetradecylphosphonic acid (PA-
C14), and n-octadecylphosphonic acid (PA-C18) were purchased from
Alfa Aesar and used without further purification. Pentacene (P5) for the
p-type organic semiconductor TFTs was commercially available and was
purified by published procedures.52 The multiple high vacuum tem-
perature gradient sublimed perylene diimide semiconductor (PDIF-
CN2) from Polyera Corp. was chosen as the n-type organic semicon-
ductor. The precursors for zinc tin oxide were prepared using zinc
acetate dihydrate (98%, Aldrich) and tin chloride(II) (98%, Aldrich).
Synthesis of Phosphonic Acid-Based π-Electron Precursor
PAE (5). Synthesis of Diethyl p-Tolylphosphonate (2). A solution of
1-bromo-4-methylbenzene (3.42 g, 20 mmol) in 140 mL of anhydrous
THF was stirred at ꢀ78 °C under nitrogen, and 1 equiv of n-BuLi in hexane
(1.6 M) was added. After 10 min, diethylchlorophosphate (3.45 g, 20 mmol)
was added to the reaction mixture, and then stirring was continued for 1 h.
After reaching room temperature, ethyl ether (100 mL) and aqueous
NaHCO3 (∼1 g in 100 mL of H2O) were added for extraction. The organic
layer was then separated and concentrated. 1H NMR (500 MHz, CD3Cl): δ
7.70ꢀ7.74 (m, 2H), 7.28ꢀ7.30 (m, 2H), 4.06ꢀ4.15 (m, 4H), 2.42 (s, 3H),
1.31ꢀ1.34 (t, 6H).
6.5 Hz), 7.90 (d, 2H, JHꢀH = 9.5 Hz), 7.70ꢀ7.74 (m, 2H), 7.55ꢀ7.57 (m,
2H), 7.10 (d, 2H, JHꢀH = 9.5 Hz), 5.04 (s, 2H), 3.65ꢀ3.74 (m, 8H). 31
P
NMR (400 MHz, DMSO-d6): δ ꢀ12.83 (s, 1P). High resolution EI-MS
calcd for C22H26N4O5Pþ: 457.16408. Found: 457.16532.
Fabrication and Characterization of Zirconia-Phosphonate
Self-Assembled Nanodielectric (Zr-SAND) Films. Zr-SAND Film
Fabrication. Heavily n-doped silicon (Montco Silicon Tech, Inc.) substrates
were cleaned in EtOH (Aldrich, absolute, 200 proof) by sonication for 2
min and then dried under flowing nitrogen, followed by oxygen plasma
treatment for 5 min to remove organic contamination and to improve the
wettability. A 0.01 M absolute ethanol solution of zirconium(IV) chloride
was prepared, followed by the addition of a mixture of nitric acid and DI
water (molar ratio; ZrCl4:HNO3:H2O = 1:10:10). The resulting zirconium
precursor solution was then heated to 50 °C for 3 h to accelerate hydrolysis.
For the ultrathin ZrO2 primer film fabrication, the hydroxyl-functionalized
Si/SiO2 substrates were spin-coated with the zirconium precursor solution
at 5000 rpm for 30 s, and then cured at 150 °C for 20 min. This procedure
was repeated twice to complete the ZrO2 primer layer. For deposition of the
first phosphonic-acid based π-electron (PAE) layer, the substrate with the
primer layer was immersed in a solution of the phosphonic-acid based
π-electron precursor (PAE; 3.0 mM in methanol) at 60 °C for 30 min. After
rinsing with MeOH, the samples were the dried under a nitrogen stream.
The third step of the growth process was accomplished by again spin-
coating the zirconium precursor, followed by curing at 150 °C for 20 min to
form a ZrO2 interlayer. Repeating the second and third steps results in
thicker multilayer formation. All zirconium precursor solutions were filtered
through a 0.2 μm pore size PTFE membrane syringe filter prior to spin-
coating and were spin-coated under a controlled atmosphere of less than
10% relative humidity (measured with a Fisher Scientific Traceable
Hygrometer-Thermometer-Dew Point probe). For MIS structures, gold
electrodes were directly deposited (200 μm ꢁ 200 μm) onto each Zr-
SAND specimen through a shadow mask.
Synthesis of Diethyl 4-(Bromomethyl)phenylphosphonate (3). A small
amount of azobisisobutyronitrile (0.016 g, 0.10 mmol) was added in
portions to a suspension of diethyl p-tolylphosphonate (0.73 g, 3.2 mmol)
and N-bromosuccinimide (0.57 g, 3.2 mmol) in anhydrous CCl4 (30 mL).
The mixture was refluxed for 4 h, then cooled to 10 °C, and filtered to
remove succinimide. The filtrate was washed with water (30 mL) and brine
(30 mL) and then dried over anhydrous Na2SO4. After filtration, the crude
product was obtained via evaporation of the solvent and was purified by
chromatography on silica gel (hexene/ethyl acetate: 2/1 as eluent) to
provide diethyl 4-(bromomethyl)phenylphosphonate as a colorless oil. 1H
NMR (500 MHz, CD3Cl): δ 7.79ꢀ7.83 (m, 2H), 7.50ꢀ7.52 (m, 2H),
4.51 (s, 2H), 4.09ꢀ4.18 (m, 4H), 1.33ꢀ1.36 (t, 6H).
Zr-SAND Film Microstructure Characterization. Optical absorption
spectra of Zr-SAND films deposited on glass were acquired with a Varian
Cary 5E spectrophotometer. The film thicknesses of Zr-SAND films on
Si wafers were analyzed by X-ray reflectivity (XRR) using wavelength λ =
0.1541 nm Cu KR radiation on an 18 kW Rigaku ATX-G diffractometer at
Northwestern University’s X-ray Diffraction Facility. The structural details
of the electron density profile are obtained by fitting the XRR data to
a multilayer model calculated by the Abeles matrix method using the
Motofit software package.53,54 Each bilayer, n, is defined by its thickness, dn,
electron density, Fn, and interface roughness, σn, to generate a complete
Synthesis of 4-[[4-[Bis(2-hydroxyethyl)amino]phenyl]diazenyl]-1-
(4-phosphonobenzyl) Pyridinium Bromide (4). A mixture of 4-[[4-
[N,N-bis(hydroxylethyl)amino]phenyl]azo]pyridine (1)47 (0.286 g,
10241
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
Scheme 2. Fabrication Procedure for Zr-Phosphonate Self-Assembled Nanodielectric (Zr-SAND) Multilayers
electron density profile F(z). The zirconia-phosphonate self-assembled
nanodielectrics are referred to in the rest of this Article as Zr-SANDs. Four
kinds of multilayer systems were prepared and analyzed: Zr-SAND-n,
where n is the number of bilayers and n = 1, 2, 3, and 4. Zr-SAND-1 was
modeled as 3 slabs (primer ZrO2 layer, PAE layer, and ZrO2 interlayer)
with a Si substrate and an air superstrate. For thicker multilayers, two
slabs were added to model each additional layer, corresponding to the
organic and inorganic components. Interparameter constraints and
batch-fitting procedures were employed to reduce the number of free
parameters, limit solutions to a physically reasonable range, and enforce
structural consistency between multilayer models. Cross-sectional TEM
samples were prepared and imaged using a JEOL-2100F scanning/
transmission electron microscope (S/TEM) and a Hitachi HD-2300A
scanning electron microscope (STEM), with both bright-field (BF) and
high-angle annular dark-field (HAADF) detectors. The morphologies of
all thin films were evaluated by atomic force microscopy (AFM) using a
JEOL-5200 scanning probe microscope with silicon cantilevers operat-
ing in the tapping mode.
Fabrication of Thin-Film Transistors Using Organic/Inorganic
Semiconductors on Zr-SAND Films. Pentacene (p-type Organic
Semiconductor) Thin-Film Transistor Fabrication. Bottom gate/top contact
OTFTs were fabricated by vacuum deposition of pentacene (50 nm, 5 ꢁ
10ꢀ6 Torr 0.05 nm/s) onto Zr-SANDs having four different thicknesses (Zr-
SAND-1, 2, 3, and 4). To complete the TFT structure, gold S/D electrodes
were vacuum-deposited (50 nm, 0.02 nm/s) through a shadow mask (L =
100 μm, W = 2000 μm).
Organic Thin-Film Transistor Fabrication (p- and n-type) and
Characterization on PA-SAM-treated Zr-SAND Films. After depositing
various PA-SAMs (PA-C6, PA-C12, PA-C14, and PA-C18) on Zr-SAND-3,
advancing aqueous contact angles were measured with DI. For OTFT
structures, bottom gate/top contact OTFTs were fabricated by vacuum
deposition of pentacene (50 nm, 5 ꢁ 10ꢀ6 Torr 0.05 nm/s) as a p-type and
PDIF-CN2 (50 nm, 5 ꢁ 10ꢀ6 Torr 0.03 nm/s) as an n-type semiconductor
onto each PA-SAMs treated Zr-SAND-3. To complete the OTFT structure,
gold S/D electrodes were vacuum-deposited (50 nm, 0.02 nm/s) through
a shadow mask (L = 100 μm, W = 2000 μm). The morphologies of
the various pentacene films were evaluated by atomic force microscopy
(AFM) using a JEOL-5200 scanning probe microscope with silicon
cantilevers operating in the tapping mode. X-ray diffraction (XRD) patterns
of pentacene thin-film (50 nm) were measured on a Rigaku ATX-G
Thin-Film Diffraction Workstation using Cu KR radiation.
TFT Measurements and Characterization of Dielectrics and Thin-
Film Transistors. MIS direct current measurements and OTFT mea-
surements were carried out under ambient conditions using a Signatone
probestation interfaced to a Keithley 6430 Sub-Femtoamp Remote
Source Meter and a Keithley 2400 source meter with a locally written
LabVIEW program. An impedance analyzer (HP 4192A) was used for
capacitance measurements.
’ RESULTS AND DISCUSSION
Here, we present a detailed investigation of the properties of
zirconia-phosphonate self-assembled nanodielectrics (Zr-SANDs)
specifically designed to form multilayer structures incorporating
highly polarizable phosphonic acid-based π-electron (PAE) layers
interleaved with ultrathin high-k zirconium oxide (ZrO2) layers. To
verify that structurally well-defined building blocks can be manipu-
lated for multilayer fabrication, Zr-SANDs are first examined by a
full complement of microstructural characterization techniques.
Following this, experimental dielectric properties such as electrical
insulating and capacitance properties as well as thermal stability
during high temperature (400 °C) anneal processing are also
presented. Next, the TFT response with representative organic
and inorganic semiconductors is examined. Finally, correlations
between the OTFT response and the structures of self-assembled
monolayers (SAMs) deposited directly on the Zr-SAND surface are
examined, and comparisons made. Importantly, it will be seen that
this deposition technique results in smooth, robust films at low
processing temperatures (∼150 °C), and the films exhibit excellent
electrical insulating and dielectric properties. These Zr-SANDs are
compatible with n-, p-type and organic, inorganic semiconductors
and enable significantly reduced TFT operating voltages as com-
pared to conventional SiO2 dielectrics.
Zinc Tin Oxide (n-type Inorganic Semiconductor) Thin-Film Transistor
Fabrication. For inorganic TFTs, the precursor solution was prepared as
follows. Zinc acetate dihydrate (0.3 M) and tin chloride(II) (0.3 M) in a 4:6
molar ratio were dissolved in 1 mL of 2-methoxyethanol (99%, Aldrich) in a
2.5 mL vial. To this solution, ethanolamine (0.3 mmol) as a stabilizer was
added to the vial, and the resulting clear solutions were stirred for 30 min at
room temperature before spin-coating. The zinc tin oxide precursor solution
was then spin-coated (3000 rpm, 30 s) onto the Zr-SAND substrates.
Subsequently, the spin-coated films were annealed on the hot plate at the
desired temperature (400 °C) for 5 min. To complete the TFT structure,
aluminum S/D electrodes were vacuum-deposited (30 nm, 1 nm/s)
through a shadow mask (L = 100 μm, W = 2000 μm).
Fabrication of Organic Thin-Film Transistors on Phosphonic
Acid Self-Assembled Monolayer (PA-SAMs)-Treated Zr-SAND
Film. Phosphonic Acid Self-Assembled Monolayer (PA-SAM) Prepara-
tion on Zr-SAND. The Zr-SAND-3 film was prepared by the procedures
described above. Four phosphonic acids having different alkyl chain lengths
were selected and were as follows: n-hexyl (PA-C6), n-dodecyl (PA-C12),
n-tetradecyl (PA-C14), and n-octadecyl (PA-C18) phosphonic acid
(Figure 12). Phosphonic acid self-assembled monolayers (PA-SAMs) were
prepared by immersing the substrates into solutions of 2.0 mM of PA-C6,
PA-C12, PA-C14, and PA-C18 (Alfa Aesar) in absolute EtOH. The
substrates were kept in the phosphonic acid solutions overnight, followed
by rinsing with EtOH, and drying under a nitrogen stream.
Fabrication and Characterization of Zirconia-Phosphonate
Self-Assembled Nanodielectric (Zr-SAND) Films. The phosphonic
10242
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
Figure 2. (A) Optical spectroscopic data as a function of reaction/
deposition time for optimizing the formation of a chromophore mono-
layer. (B) UVꢀvis absorption spectra of Zr-SAND-1 (black), Zr-SAND-
2 (red), Zr-SAND-3 (blue), and Zr-SAND-4 (green) samples. (C)
Optical absorption spectra of Zr-SAND-n at 575 nm as a function of the
number of bilayers (Zr-SAND-1, 2, 3, and 4).
Figure 3. (A) X-ray reflectivity data and theoretical fit as a function of q
for Zr-SAND-n multilayer films with number of bilayers n = 1 (black),
n = 2 (red), n = 3 (blue), and n = 4 (green). (B) The electron density
profiles (normalized to FSi), corresponding to the fits in (A) as a function
of height z above the Si substrate surface.
acid-based π-electron building block (PAE) 4-((4-(bis(2-hydro-
xyethyl)amino)phenyl)diazenyl)-1-(4-(diethoxyphosphoryl)benzyl)
pyridinium bromide (5; Scheme 1) was synthesized as described in
the Experimental Section and was characterized by conventional
analytical techniques (1H NMR, 13C NMR, 31P NMR, and HR-MS).
The zirconia-phosphonate self-assembled nanodielectrics (Zr-
SANDs) were fabricated according to Scheme 2, consisting of
depositing the ZrO2 primer layer on a silicon wafer, followed by
iterative deposition of the phosphonic acid-based π-electron (PAE)
layer and the ZrO2 interlayer for capping and multifunction. The
ultrathin high-k ZrO2 primer layer and interlayer bonding enhance
the orientational stability of the PAE molecules and minimize
capacitance degradation of the parallel multilayer structure while
enhancing leakage characteristics. Thus, pretreated heavily n-doped
silicon was spin-coated using the zirconium precursor solution
described above, and then cured at 150 °C for 20 min to complete
the primer layer. For Zr-SAND-1 fabrication, the primer ZrO2 layer-
coated substrate was immersed in a 3 mM methanol solution of the
phosphonic acid-based π-electron (PAE) molecule at 60 °C for
30 min, rinsed with methanol, and the ZrO2 interlayer was then
deposited by the same method used for the primer layer. Zr-SAND-2,
3, and 4 structures were formed by alternating repetition of the PAE
and ultrathin ZrO2 bilayer depositions. PAE solutions can be reused
multiple times without any noticeable detrimental effects on film
quality.
The present iterative two-step process (PAE þ ZrO2 layer)
(Scheme 2) and the resulting multilayer structural regularity
were characterized by a full complement of physicochemical
techniques: (1) transmission optical spectroscopy (UVꢀvis) to
characterize assembly chemistry and microstructural regularity,
(2) X-ray reflectivity (XRR) to characterize film thickness,
density, and interfacial roughness, (3) transmission electron
microscopy (TEM) to characterize film thickness and micro-
structural regularity, and (4) atomic force microscopy (AFM) to
characterize surface morphology and roughness.
Figure 4. Specular X-ray reflectivity (XRR)-derived film thickness
(nm) data as a function of the number of bilayers, n, in the multilayer
films Zr-SAND-1ꢀ4 as prepared via the procedure of Scheme 2. The
solid line is the fit by linear regression, indicating an average bilayer
thickness of 2.4 nm.
575 nm on Zr-SAND-1ꢀ4, demonstrating that essentially equal
quantities of uniformly aligned chromophore units are incorpo-
rated in each sublayer up to Zr-SAND-4 bilayers.
The XRR data and analysis are shown in Figure 3A as a
function of the out-of-plane scattering vector q = 4π sin θ/λ, with
least-squares best-fit models calculated with Abeles matrix meth-
od using the Motofit software package (Nelson).53,54 The data
exhibit the classical Kiessig fringes expected from similar orga-
nicꢀinorganic hybrid multilayers with alternating low- and high-
electron density layers.55 The corresponding fits for each data set
are shown overlaid on the data. Easily observable is the emer-
gence of the first-order multilayer Bragg peak at q = ∼2.65 nmꢀ1
(d = ∼2.35 nm) as the number of layers increases. In Figure 3B,
the extracted electron density profiles, normalized to FSi, reveal
highly ordered multilayers, which are consistent from sample to
sample in terms of electron density, roughness, and thickness.
Well-defined ZrO2/PAE layer progressions are evident from the
strong oscillations in the electron density profile. Additionally, all
of the ZrO2 and PAE interfaces are at comparable positions in z
from sample to sample, indicating highly controlled, sequential
deposition. The total thicknesses of the Zr-SAND-1, 2, 3, and 4
films are found to be 4.7 ( 0.1, 6.7 ( 0.1, 9.5 ( 0.2, and 11.3 (
0.8 nm, respectively, as derived from the individual electron
density profiles themselves,56 and increase linearly with increas-
ing numbers of bilayers, n (Figure 4). A linear fit of the thickness
as a function of n linear regression reveals an average bilayer
(PAE þ ZrO2 interlayer) thickness of 2.4 ( 0.1 nm, with
dPAE:dZrO2 ∼1.6:1 and the ZrO2 primer layer thickness of
2.1 ( 0.2 nm. The electron density minima in the electron
The PAE deposition kinetics from MeOH solution onto sub-
strates were monitored by UVꢀvis spectroscopy to optimize
reaction conditions. Figure 2A shows the deposition kinetics at
60 °C. The PAE optical transition at 575 nm reaches a maximum
after 30 min at 60 °C, and longer reaction times result in a constant
absorption, meaning that a densely packed molecular assembly of
PAE is achieved. No other bands or shifts in the optical absorbance
maxima are observed, arguing against significant chromophore aggre-
gation or decomposition (Figure 2B). Furthermore, the UVꢀvis
measurements unambiguously demonstrate a linear dependence
of the HOMOꢀLUMO CT chromophore optical absorbance at
10243
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
Figure 7. (A) Leakage current density versus voltage (J vs V) and (B)
leakage current density versus electric field (J vs E) for Zr-SAND-1ꢀ4.
Figure 5. Cross-section TEM image of the Si/Zr-SAND-4 film inter-
face. Labels indicate the layer identities.
agreement with the aforementioned XRR results. Atomic force
microscopy (AFM) images (Figure 6) of the Zr-SAND-1ꢀ4 films
demonstrate continuous, crack/pinhole-free surface morphologies
(maximum root-mean-square (rms) roughness of ∼0.5 nm).
To assess the dielectric properties of the Zr-SANDs, metalꢀ
insulatorꢀsemiconductor (MIS) sandwich structure devices were
fabricated by thermal Au electrode deposition (dimension: 200 μmꢁ
200 μm) on the Zr-SANDs-coated nþþ-Si substrates. Figure 7
shows typical current density versus voltage (JꢀV) and current
density versus electric field (JꢀE) plots for MIS structures
fabricated with four different thicknesses (Zr-SAND-1, 2, 3,
and 4). The leakage current density of these films progressively
decreases for the same applied voltage ((4 V) as the number
of layers is increased due to the increased thickness. Note,
however, that the leakage current density for the same electric
field increases with the number of layers, then saturates at
∼10ꢀ7 A/cm2 at 2 MV/cm beyond Zr-SAND-3. The lower
leakage current density for Zr-SAND-1 or -2 versus the thicker
films appears to arise from synergistic effects involving additional
leakage barriers such as those provided by the native oxide and
the ZrO2 primer layer. However, note that this effect is reduced
with increasing layer numbers because the contribution of the
additional leakage barrier becomes negligible as the dielectric
thickness is increased. Furthermore, the leakage current densities
of the combined PAEþZrO2 hybrid films are at least 10ꢁ lower
(<1ꢁ 10ꢀ6 A/cm2, 2 MV/cm) than that of bulk ZrO2 films (∼1 ꢁ
10ꢀ5 A/cm2, 2 MV/cm) prepared by solꢀgel methods at the
same processing temperature (150 °C).58 All of these data are
summarized in Table 1.
Figure 6. AFM images at 5 ꢁ 5 μm scan area of Zr-SAND-1ꢀ4 films.
The space bar indicates 1 μm, and F is rms roughness.
density profiles correspond to the organic PAE layer and are
found to be approximately that of bulk Si; however, this value is
artificially high because the interface roughnesses are on the
order of the PAE layer thickness. The ZrO2 layers have electron
densities that are nearly twice that of bulk Si, but again it is
possible to see the effects of the interface roughness in the
reduced electron densities of the thinner ZrO2 layers deposited
on the PAE as compared to that of the ZrO2 base layer. Finally,
interfaces were found to have rms roughnesses of less than
0.5 nm from the data fitting. A cross-sectional high-angle annular
dark-field scanning transmission electron microscopy (HAADF
STEM) image of Zr-SAND-4 (Figure 5) reveals a continuous
and uniform contrast for the individual phases, which provides
direct observation of the superlattice structures and confirms
the expectations for the individual PAE and ZrO2 layers in the
multilayer film structure. The primer ZrO2 is ∼2 nm thick on top
of native Si oxide layer (as expected by the double coating), and
the PAE layer thickness is 1.5 ( 0.1 nm, in agreement with
the computed molecular lengths and previous results.48,57 The
subsequent top ZrO2 layer is ∼1 nm thick. The thicknesses of the
Zr-SANDs can be controlled by adjusting the number of PAE
and ultrathin ZrO2 bilayers. Note that the measured thickness of
∼12 nm for Zr-SAND-4 from the TEM images is in good
Capacitanceꢀfrequency (Cꢀf) measurements were per-
formed on the MIS structures, and the measured capacitances
are 750 (Zr-SAND-1), 633 (Zr-SAND-2), 535 (Zr-SAND-3),
and 465 nF/cm2 (Zr-SAND-4) at 10 kHz (Figure 8). As the
number of layers is increased, the capacitance values decrease due
to the increased thickness. The multilayer dielectric in the
n
þ2-Si/Zr-SAND [primer ZrO2/(PAE/interlayer ZrO2)n, n
(the number of bilayer) = 1, 2, 3, and 4]/Au MIS devices can
be modeled as capacitors in series, according to eq 3.
!
ꢀ ꢁ
ꢀ
ꢁ
1
1
1
1
1
¼
þ
þ n
þ
ð3Þ
3
Ci
CSiO
Cp-ZrO
CPAE Ci-ZrO
2
2
2
where n is the number of bilayers, SiO2 is the native oxide on the
Si wafer, p-ZrO2 is the ZrO2 primer layer, i-ZrO2 is the ZrO2
interlayer, and PAE is the phosphonic-acid-based π-electron
layer. The capacitances of the native oxide and ZrO2 primer
layer on the Si bottom electrode are 1380 and 4425 nF/cm2,
respectively, assuming a 2.5 nm thick SiO2 (k = 3.9) and 2 nm
10244
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
Table 1. Summary of Film and Dielectric Properties for Zr-SAND-1ꢀ4 Films
thickness (nm)
roughness (nm)
J (A/cm2) at 2 MV/cm
Ci (nF/cm2)
Zr-SAND-1
Zr-SAND-2
Zr-SAND-3
Zr-SAND-4
4.7 ( 0.1
6.7 ( 0.1
9.5 ( 0.2
11.3 ( 0.8
0.15
0.25
0.36
0.40
7 ꢁ 10ꢀ8
2 ꢁ 10ꢀ7
3 ꢁ 10ꢀ7
3 ꢁ 10ꢀ7
750
633
535
465
Figure 8. (A) Capacitance versus voltage plots measured at 10 kHz, and
(B) inverse capacitance versus number of Zr-SAND bilayers. Inset:
Capacitance versus frequency plots measured at 3.0 V forZr-SAND-1ꢀ4.
Figure 10. (A) X-ray reflectivity (XRR)-derived film thickness and (B)
inverse capacitance versus number of Zr-SAND bilayers for the control
and thermal annealed Zr-SAND multilayer samples.
the same bias window (ꢀ4 to þ4 V). As compared to the control
Zr-SAND-4 film (∼6 ꢁ 10ꢀ6 A/cm2 at 4 V), the leakage current
density increases slightly at the same applied voltage, after
annealing at 400 °C in air (∼3 ꢁ 10ꢀ5 A/cm2 at 4 V, Figure
S1A). The capacitance of the Zr-SAND-4 film likewise increases
as the annealing temperature is increased, from 465 (control) to
650 nF/cm2 (400 °C; Figure S2). To further understand leakage
current and capacitance variation during annealing, the micro-
structure of Zr-SAND-4 before annealing at high temperature
and after thermally annealing at 400 °C was investigated by X-ray
reflectivity (XRR). The XRR data and the extracted electron
density profiles for control Zr-SANDs and high-temperature
annealed Zr-SANDs are shown in Figure 9, with the best fits. The
analysis of the XRR data reveals that the highly ordered multi-
layer structure persists after thermal annealing at 400 °C. How-
ever, as compared to the control Zr-SAND films, the thicknesses
are reduced upon thermal annealing by as much as 20% versus the
control Zr-SAND film (Figure 10A). This contraction is similar to
that in silane-based conventional SAND dielectric films.59 Note
that while the film becomes thinner and denser after thermal
annealing, the microstructural data and reciprocal Ciꢀ1 versus Zr-
SAND-n (n = 1, 2, 3, and 4) plot (Figure 10B) demonstrates that
well-defined superlattice structures are preserved in the high
temperature annealing. Therefore, the increased capacitance is
reasonably associated with the decreased Zr-SAND film thickness.
Note that the leakage current density increases slightly after
400 °C annealing for the same bias window and may simply reflect
the film thickness reduction. Figure S1B plots the leakage current
density versus electric field, demonstrating the similar behavior of
the leakage current density at the same electric field between control
and thermal annealed Zr-SAND-4 film. These results provide clear
evidence that nanoscopic SAND films exhibit remarkable thermal
and dielectric stability under ambient, rendering them suitable for
high-temperature film growth/annealing processes.
Figure 9. (A) X-ray reflectivity data and theoretical fits for control and
thermally annealed Zr-SANDmultilayers as a function of q. (B) The
electron density profiles (normalized to FSi), corresponding to the fits in
(A) as a function of height z above the Si substrate surface.
thick ZrO2 layer (k = 10),58 respectively. From the accumulation
regime capacitances, the Ci of each Zr-SAND is measured as
465ꢀ750 nF/cm2 at 10 kHz, yielding a PAE dielectric constant
of ∼7 for the entire multilayer structure. The reciprocal value of
Ci versus Zr-SAND-n (n = 1, 2, 3, and 4) plot shows a linear
increase with the number of bilayers, which supports a regularly
defined multilayer structure.
Thermal Stability of the Zr-SAND under High Annealing
Temperature Conditions. Prior to Zr-SAND-based inorganic
TFT fabrication, the thermal stability of the Zr-SAND dielectrics
under ZTO annealing conditions was investigated in MIS
capacitor structures. The nþ-Si/Zr-SAND substrates were ther-
mally annealed at 400 °C, which is the same temperature as ZTO
thin films for 5 min in air, and Au dot contacts then thermally
evaporated through a shadow mask (200 ꢁ 200 mm2). Next, the
leakage current density (Jleak) versus voltage (or electric field)
and capacitance versus voltage (or frequency) of the control Zr-
SAND-4 and thermally annealed Zr-SAND-4 were measured in
Fabrication and Characterization of Thin-Film Transistors
Using Organic or Inorganic Semiconductors on Zirconia-
Phosphonate Self-Assembled Nanodielectric (Zr-SAND)
Films. Given the excellent dielectric properties of Zr-SANDs,
TFTswerenextfabricatedwithrepresentative organic (pentacene;
P5) and inorganic (zinc tin oxide; ZTO) semiconductors on
10245
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
Figure 12. (A) Schematic view of a top contact OTFT using different n-
alkyl PA-SAM/Zr-SAND gate dielectrics. (B) Chemical structures of the
n-alkyl PA-SAMs used in this study: n-hexylphosphonic acid (PA-C6), n-
dodecylphosphonic acid (PA-C12), n-tetradecylphosphonic acid (PA-
C14), n-octadecylphosphonic acid (PA-C18).
Figure 11. (A) Transfer plots and (B) output plots for pentacene
OTFTs. (C) Transfer plots and (D) output plots for ZTO TFTs based
on Zr-SAND-4.
Table 2. Carrier Mobility (μ, cm2 Vꢀ1 sꢀ1), Threshold
Voltage (Vt), and Current On/Off Ratio (Ion/Ioff) Data for
TFT Devices Fabricated Using Pentacene and Zinc Tin Oxide
as the p-type Organic and n-type Inorganic Semiconductors
on Zr-SAND-1ꢀ4 Films
P5
ZTO
μ
μ
(cm2 Vꢀ1 sꢀ1
)
Vt Ion/Ioff (cm2 Vꢀ1 sꢀ1
)
Vt Ion/Ioff
Zr-SAND-1 0.32 ( 0.02 ꢀ0.5 104
Zr-SAND-2 0.35 ( 0.02 ꢀ0.7 105
Zr-SAND-3 0.36 ( 0.01 ꢀ0.8 105
Zr-SAND-4 0.38 ( 0.01 ꢀ0.9 105
3.0 ( 0.1
3.3 ( 0.1
3.4 ( 0.1
3.5 ( 0.1
0.9 105
1.0 106
1.1 107
1.1 107
Figure 13. Variation in (A) aqueous contact angle (deg) and (B)
capacitance as a function of the number of carbon atoms in the n-alkyl
phosphonic acid SAMs deposited on Zr-SAND-3 films.
((100 V),60ꢀ63 reflecting the greater capacitance of the Zr-
SANDs versus a conventional SiO2 gate dielectric.
Zr-SANDs by vacuum vapor deposition and spin-coating, respec-
tively. All devices exhibit reproducible IꢀV characteristics at low
bias (<(4 V). Typical IꢀV plots for TFTs on Zr-SAND-4 are
shown in Figure 11, and Table 2 summarizes performance
parameters asa functionofsemiconductoron Zr-SAND dielectrics
having various thicknesses. Using the above capacitances and
device geometry (L = 100 μm, and W = 2000 μm), these P5
OTFTs devices exhibit good performance with hole mobilities of
0.35ꢀ0.38 cm2 Vꢀ1 sꢀ1, low threshold voltages (ꢀ0.5 to ꢀ0.9 V),
and onꢀoff current ratios of 104ꢀ105 on Zr-SAND-1ꢀ4. The
ZTO TFTs fabricated with Zr-SAND exhibit excellent IꢀV
characteristics with classical/crisp linear pinch-off curves and
saturation at very low operating voltages (<4.0 V). Analysis of
the ZTO-TFT electrical response reveals large saturation-regime
Fabrication and Characterization of Organic Thin-Film
Transistors on Phosphonic-Acid Self-Assembled Monolayer
(PA-SAM)-Functionalized Zr-SANDs. The present zirconia-
phosphonate self-assembled nanodielectrics (Zr-SANDs) have a
ZrO2 surface which is decidedly hydrophilic (advancing aqueous
contact angle <10°). Note that this characteristic of these high-k
inorganic materials negatively affects organic semiconductor
(OSC)-based TFT device performance through a combination
of charge carrier surface trap sites associated with ꢀOH groups,
induced ionic polarization between charge carriers, and the high-k
ionic lattice.64ꢀ66 To address these limitations, such high-k oxide
surfaces are typically planarized with a polymer dielectric leaving a
nonpolar low-k surface, which is generally more compatible with
standard OSCs;67ꢀ69 however, the addition of a thicker polymer
film reduces the capacitance. Another promising approach is to
deposit nanoscopic self-assembled monolayers (SAMs), which are
molecular films that spontaneously form by the adsorption of an
organic surfactant on a solid substrate from the gas or solution
field-effect mobilities of 3.0ꢀ3.5 cm2 Vꢀ1
s
ꢀ1, and excellent
onꢀoff current ratios of 105ꢀ107. These carrier mobilities are
comparable to or greater than those of control devices fabricated
with a conventional 300 nm thick SiO2 gate dielectric (mobility
∼0.26 and 1.8 cm2 Vꢀ1 sꢀ1 for P5 and ZTO, respectively), and
their operating voltages (<(4 V) are much lower than on SiO2
10246
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
Table 3. Carrier Mobility (μ, cm2 Vꢀ1 sꢀ1), Threshold
Voltage (Vt), and Current On/Off Ratio (Ion/Ioff) Data for
TFT Devices Fabricated Using Pentacene and PDIF-CN2 as
Representative p- and n-type Organic Semiconductors on
PA-SAMs/Zr-SAND-3
P5
Vt
PDIF-CN2
Vt Ion/Ioff
CA Ci
μ
Ion/Ioff
μ
Zr-SAND-3 <10 560 0.36 ꢀ1.0
105
105
106
106
105
0.36 ꢀ1.2
0.41 ꢀ1.0
102
103
104
104
104
PA-C6
89 465 0.61 ꢀ0.7
105 405 0.74 ꢀ0.6
105 398 0.68 ꢀ0.8
107 365 0.16 ꢀ0.5
PA-C12
PA-C14
PA-C18
0.63
0.79
0.64
0.02
1/2
Figure 14. (A) Transfer characteristics and (B) IDS ꢀVG plots of P5
TFTs fabricated on Zr-SAND-3 capped with the indicated PA-SAMs.
The gate voltage is swept at a constant drainꢀsource voltage VDS = ꢀ3 V.
Inset: Chemical structure of pentacene (P5).
0.01
0.3
mostly disordered films, possibly due to a lack of cohesive van der
Waals interactions between the chains.74,75 For SAM chains longer
than about 10 carbon atoms, the cohesive forces are expected to be
sufficiently strong to force the molecules into an almost upright
position, yielding dense, well-ordered monolayers, and affording
large aqueous contact angles (>105°).74ꢀ77 Note, however, that
the surface properties of all of the present PA-SAM functionalized
Zr-SANDs are hydrophobic. The leakage current densities of the
PA-SAM/Zr-SAND-3 films (Figure S3) are not significantly
reduced at the same applied voltage ((4 V). However, the Ci of
the PA-SAM/Zr-SAND-3 films (Figure 13B) decreases with
increased SAM alkyl chain length from 560 nF/cm2 for the bare
Zr-SAND-3 film to 365 nF/cm2 for PA-C18 due to the relatively
low dielectric constant of the hydrocarbon chains.28 AFM was used
to characterize the morphology of the n-alkyl PA-SAMs assembled
on Zr-SAND, revealing generally smooth films (rms roughness less
than 0.5 nm) (Figure S4). Figures 14 and 15 show representative
transfer plots for OTFTs fabricated with P5 and PDIF-CN2 on PA-
SAMs/Zr- SAND-3. These plots exhibit reproducible IꢀV char-
acteristics at low operating voltages (<(4.0 V) as well as excellent
linear/saturation behavior. For the P5 TFTs, Figure 14 shows that
increasing the alkyl chain length from 6 to 14 carbon atoms causes
the mobility to increase from 0.61 to 0.74 cm2 Vꢀ1 sꢀ1, while further
increasing the chain length to 18 carbon atoms causes the mobility
to drop to 0.16 cm2 Vꢀ1 sꢀ1. Note that the P5 mobility on the bare
Zr-SAND is only 0.36 cm2 Vꢀ1 sꢀ1 and therefore smaller than most
of the OTFTs fabricated with PA-SAM functionalized Zr-SAND.
All of these devices exhibit excellent onꢀoff current ratios (up to
105). Figure 14B and Table 3 show that the threshold voltages (Vt)
decrease from ꢀ1.0 V for P5-TFTs on bare Zr-SAND to ꢀ0.5 V for
TFTs with PA-SAMs. The threshold voltage is closely related to
the charge state at the organic semiconductor channel/gate di-
electric interface as long as the entire semiconductor channel films
contain about the same density of deep level defects.66,78 There-
fore, the threshold voltage depends strongly on the preparation of
the surface on which the organic semiconductor is deposited. In
general, SAM deposition before growing the organic semiconduc-
tor layers is likely to eliminate trap states on the dielectric
surface.72,73,78,79 Consequently, as compared to bare Zr-SAND
substrate, the PA-SAM functionalized surfaces should provide
lower trap state densities, resulting in lower threshold voltages and
higher TFT mobilities.80
1/2
Figure 15. (A) Transfer characteristics and (B) IDS ꢀVG plots of
PDIF-CN2 TFTs fabricated on Zr-SAND-3 coated with the indicated
PA-SAMs. The gate voltage is swept at a constant drainꢀsource voltage
V
DS = 3 V. Inset: Chemical structure of PDIF-CN2.
phase.70,71 With the addition of a few nanometers thick SAM layer,
the dielectric capacitance is not reduced as much as by a conven-
tional polymer film. In addition, functional SAMs provide an
effective method for controlling the dielectric surface properties
that affect the growth of the overlying thin film crystal structure
and morphology, and hence the OSC electrical properties and
subsequent OTFT device performance.28,32,72,73 Here we system-
atically investigate the relationship between the alkyl chain
length of the phosphonic-acid SAMs (PA-SAMs) that coat the
gate dielectric and the electrical performance characteristics of
the resulting organic TFTs (Figure 12). Commercially available
phosphonic acids with four different alkyl chain lengths,
CH3(CH2)xPO(OH)2, x = 5, 11, 13, and 17, were selected and
compared. The PA-SAMs were prepared by immersing substrates
coated with Zr-SAND-3 in an absolute EtOH solution of the
corresponding phosphonic acid at room temperature for 12 h.
Next, the substrates were rinsed with EtOH and dried in a stream
of nitrogen. The purified semiconductors pentacene (P5) and
perylene diimide (PDIF-CN2) were then vacuum-deposited to
form 50 nm thick semiconductor layers on the PA-SAM-treated
dielectric. Advancing aqueous contact-angle measurements
using DI water show that due to n-alkyl PA-SAM formation,
the ZrO2 surface changes from hydrophilic (10°) to hydrophobic
(89ꢀ107°) (Figure 13A). The slight decrease in aqueous contact
angle from 105° for PA-C12 to 89° for PA-C6 may be a result of
lower SAM densities on the substrate. SAMs derived from
hydrocarbon chains having less than about 10 carbon atoms form
Using a similar dielectric modification process, we also in-
vestigated n-type organic transistors fabricated with PDIF-CN2
(Figure 15), also operating at low voltages (<4.0 V). The TFT
electron mobilities increase as the PA alkyl chain length is
10247
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
charge transport is confined to the nanoscopic region at the
semiconductor/dielectric interface, an optimized interface be-
tween the dielectric film and the semiconductor is clearly a
prerequisite for optimum OTFT performance.60,79 Especially for
P5-based TFTs, their interfacial properties are closely related to
the morphology and microstructure of the P5 thin films, thereby
strongly influencing their performance. We have found here that
P5 TFT characteristics depend significantly on the SAM alkyl
chain lengths, and that PA-C12 or PA-C14 provide optimum
performance for the P5 semiconductor films in this study. We
have evaluated the P5 thin-film morphology by AFM and the P5
thin-film crystallinity on the various PA-SAMs/Zr-SAND-3 by
X-ray diffraction (XRD). Figure 16 compares the P5 grain size
and crystallinity of P5 films grown on bare Zr-SAND-3, and on
PA-C6, C12, C14, and C18/Zr-SAND-3. According to these
data, the grain sizes of the P5 thin films (approximately 1 μm)
decrease slightly from bare Zr-SAND-3 to PA-C12, and then drop
to ∼0.5 μm on PA-C14-coated Zr-SAND-3. Eventually, the crystal-
lite dimensions become very small (less than 0.1 μm) on the PA-
C18/Zr-SAND-3. On PA-C14/Zr-SAND-3, the P5 film has a grain
size of approximately 0.5 ( 0.1 μm, which is significantly larger than
in films on PA-C18/Zr-SAND-3. Figure 16B shows the XRD
patterns of the P5 films grown on PA-SAMs/Zr-SAND-3. The
XRD data for P5 films on PA-SAMs/Zr-SAND-3 feature a dis-
tinctive first-order reflection located at 2θ = 5.7°. Three character-
istic features arising from the (00l) Bragg planes of the thin-film
phase dominate the XRD patterns. These films on bare Zr-SAND-3
and on Zr-SAND-3 coated with PA-C6, PA-C12, and PA-C14 are
highly crystalline because the first reflection is sharp and intense. In
contrast, P5 films grown on PA-C18 are far less crystalline. After
functionalizing the Zr-SAND-3 with PA-SAMs, the device perfor-
mance generally improves even though the grain size of the P5 films
is reduced somewhat as compared to the bare Zr-SAND-3 substrate.
This performance enhancement when comparing PA-SAMs to bare
Zr-SAND-3 (Table 3) verifies that the presence of the SAMs
enhances the quality of the OSC/dielectric interface. The lower
Vt and higher μ values observed for most n-alkyl PA-SAMs versus
bare Zr-SAND-3 are consistent with the suppression of any adverse
effects of the high-k dielectric surface on device performance by
reduction of ꢀOH trap sites on the surface.66,78 When comparing
TFT results between the different n-alkyl PA-SAMs, there are
distinctive differences, particularly in mobility. Furthermore, there
are systematic trends that are observed in μ when comparing PA-
C12, PA-C14, and PA-C18/Zr-SAND-3 (Figure 17). As the
number of carbon atoms of the SAM alkyl chain is increased from
C12 to C18, the relative P5 XRD diffraction intensities decrease and
the grain sizes also decline, resulting in clear degradation of μ from
0.74 to 0.16 cm2 Vꢀ1 sꢀ1. FortheP5 thin films on PA-C18, the grain
size and XRD intensity fall significantly, resulting in very poor TFT
performance. This trend is expected because it is known that, in
general, small grain sizes and reduced film crystallinity depress μ due
to drastically increased grain boundary densities.81ꢀ84 Figure 17
shows the grain size, relative crystallinity, and mobility trends of the
P5 films grown on bare Zr-SAND-3, and on Zr-SAND-3 coated
with PA-C6, PA-C12, PA-C14, and PA-C18.
Figure 16. (A) AFM images and (B) θꢀ2θ X-ray diffraction patterns of
P5 films grown on Zr-SAND-3 treated with the indicated PA-SAMs.
AFM images are of areas 5 μm ꢁ 5 μm in size.
Figure 17. Variation in (A) pentacene film grain size, (B) normalized
intensity of the pentacene film (00l) Bragg reflection, and (C) pentacene
field-effect mobility in low-voltage OTFTs as a function of the number of
alkyl carbon atoms in the n-alkyl phosphonic acid SAM on Zr-SAND-3.
increased from C6 to C14, but then fall slightly for the longest
chain length (C18). Note that the TFT performance with the
SAM-modified Zr-SAND is significantly greater than on a bare
Zr-SAND surface. The on/off ratios are enhanced by ∼100ꢁ
with the PA-SAMs, and the threshold voltage shifts to positive
(0.01 V) from negative (ꢀ1.2 V) as the PA-SAM chain length is
increased. All of these data are summarized in Table 3.
Using the same SAM modification process, we also investi-
gated n-type organic transistors using the semiconductor PDIF-
CN2. The OTFT performance with PDIF-CN2 shows behavior
generally similar to that of the P5 TFTs (Table 3). TFT
mobilities with the SAM-modified Zr-SAND are substantially
enhanced versus those on the bare Zr-SAND surface, and on/off
ratios are enhanced by ∼100ꢁ with the PA-SAMs. AFM and
Pentacene Morphology and OTFT Response Characteris-
tics on PA-SAM-Functionalized Zr-SAND. Because OTFT
10248
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
XRD data (Figure S5) reveal similar PDIF-CN2 morphologies
for SAM-coated Zr-SAND versus bare Zr-SAND. However, in
terms of the threshold voltage, large shifts to positive (0.01 V)
from negative (ꢀ1.2 V) voltages are observed as the PA-SAM
chain length is increased. The systematic PA-SAM chain length
increases result in substantially increased the electron carrier
density in the transistor channel, inducing an abrupt shift in the
threshold voltage toward more positive values. This relationship
between the SAM-treated gate dielectric surface and the thresh-
old voltage is similar to that observed in other p- and n-type
organic TFTs with various SAM gate dielectrics.80,85
(5) Cho, J. H.; Lee, J.; Xia, Y.; Kim, B.; He, Y. Y.; Renn, M. J.; Lodge,
T. P.; Frisbie, C. D. Nat. Mater. 2008, 7, 900.
(6) Klauk, H. Organic Electronics: Materials, Manufacturing and
Applications; Wiley-VCH: Weinheim, 2006.
(7) Kagan, C. R.; Andry, P. Thin-Film Transistors; Marcel Dekker:
New York, 2003.
(8) Sze, S. M.; Ng, K. K. Physics of Semiconductor Devices, 3rd ed.;
Wiley-Interscience: Hoboken, NJ, 2007.
(9) Ramesh, R.; Schlom, D. G. MRS Bull. 2008, 33, 1006.
(10) Pal, B. N.; Dhar, B. M.; See, K. C.; Katz, H. E. Nat. Mater. 2009,
8, 898.
(11) Murphy, A. R.; Frechet, J. M. J. Chem. Rev. 2007, 107, 1066.
(12) Sekitani, T.; Zschieschang, U.; Klauk, H.; Someya, T. Nat.
Mater. 2010, 9, 1015.
(13) Loo, Y. L.; McCulloch, I. MRS Bull. 2008, 33, 653.
(14) Marks, T. J. MRS Bull. 2010, 35, 1018.
(15) Facchetti, A. Mater. Today 2007, 10, 28.
(16) Yan, H.; Chen, Z. H.; Zheng, Y.; Newman, C.; Quinn, J. R.;
Dotz, F.; Kastler, M.; Facchetti, A. Nature 2009, 457, 679.
(17) Zaumseil, J.; Sirringhaus, H. Chem. Rev. 2007, 107, 1296.
(18) Sun, Y. G.; Rogers, J. A. Adv. Mater. 2007, 19, 1897.
(19) Kim, F. S.; Guo, X. G.; Watson, M. D.; Jenekhe, S. A. Adv. Mater.
2010, 22, 478.
(20) Wong, W. W. H.; Singh, T. B.; Vak, D.; Pisula, W.; Yan, C.; Feng,
X. L.; Williams, E. L.; Chan, K. L.; Mao, Q. H.; Jones, D. J.; Ma, C. Q.;
Mullen, K.; Bauerle, P.; Holmes, A. B. Adv. Funct. Mater. 2010, 20, 927.
(21) Hutchison, G. R.; Ratner, M. A.; Marks, T. J. J. Am. Chem. Soc.
2005, 127, 2339.
(22) Letizia, J. A.; Salata, M. R.; Tribout, C. M.; Facchetti, A.; Ratner,
M. A.; Marks, T. J. J. Am. Chem. Soc. 2008, 130, 9679.
(23) Facchetti, A.; Yoon, M. H.; Stern, C. L.; Katz, H. E.; Marks, T. J.
Angew. Chem., Int. Ed. 2003, 42, 3900.
(24) Veres, J.; Ogier, S.; Lloyd, G.; de Leeuw, D. Chem. Mater. 2004,
16, 4543.
’ CONCLUSIONS
We have demonstrated here that a new solution-phase layer-
by-layer growth method for self-assembled organicꢀinorganic
multilayer dielectrics using phosphonic acid-based organic and
zirconium-based inorganic precursors affords robust, smooth,
adherent, insulating, pinhole-free, high-capacitance, thermally
stable, ultrathin dielectric materials. These multilayer materials
can be used for low operating voltage organic and inorganic
semiconductor-based TFT devices as versatile gate dielectrics
with excellent thermal stability. The advantages of the present
growth method include accurate control of film thickness, large-
area uniformity, highly conformal layering, and easily tunable
interfacial properties. Furthermore, this method is an ideal
fabrication technique for various organicꢀinorganic hybrid
multilayer structures and should also be applicable to fabricating
a variety of organicꢀinorganic hybrid superlattices, which facil-
itate novel materials processing for diverse applications.
’ ASSOCIATED CONTENT
(25) Ortiz, R. P.; Facchetti, A.; Marks, T. J. Chem. Rev. 2010, 110, 205.
(26) Facchetti, A.; Yoon, M. H.; Marks, T. J. Adv. Mater. 2005,
17, 1705.
(27) Robertson, J. Rep. Prog. Phys. 2006, 69, 327.
(28) Acton, O.; Ting, G.; Ma, H.; Ka, J. W.; Yip, H. L.; Tucker, N. M.;
Jen, A. K. Y. Adv. Mater. 2008, 20, 3697.
S
Supporting Information. AFM images and dielectric
b
performance data for selected dielectric materials. This material
is available free of charge via the Internet at http://pubs.acs.org.
(29) Costescu, R. M.; Cahill, D. G.; Fabreguette, F. H.; Sechrist,
Z. A.; George, S. M. Science 2004, 303, 989.
(30) Mitzi, D. B. Chem. Mater. 2001, 13, 3283.
(31) Klauk, H.; Zschieschang, U.; Pflaum, J.; Halik, M. Nature 2007,
445, 745.
(32) Zschieschang, U.; Ante, F.; Yamamoto, T.; Takimiya, K.;
Kuwabara, H.; Ikeda, M.; Sekitani, T.; Someya, T.; Kern, K.; Klauk, H.
Adv. Mater. 2010, 22, 982.
(33) Zschieschang, U.; Ante, F.; Schlorholz, M.; Schmidt, M.; Kern,
K.; Klauk, H. Adv. Mater. 2010, 22, 4489.
(34) Acton, O.;Osaka,I.; Ting, G.;Hutchins, D.;Ma,H.;McCullough,
R. D.; Jen, A. K. Y. Appl. Phys. Lett. 2009, 95.
(35) Ting, G. G.; Acton, O.; Ma, H.; Ka, J. W.; Jen, A. K. Y. Langmuir
2009, 25, 2140.
’ AUTHOR INFORMATION
Corresponding Author
a-facchetti@northwestern.edu; t-marks@northwestern.edu
’ ACKNOWLEDGMENT
This research was supported by the AFOSR (FA9550-08-1-
0331) and the NSF MRSEC program (DMR-0520513) at
Northwestern University. Microscopy studies were performed
in the EPIC and NIFTI facilities of the NUANCE Center at
Northwestern University. The NUANCE Center is supported by
NSF-NSEC, NSF-MRSEC, the State of Illinois, and North-
western University. This work made use of the J. B. Cohen
X-ray Diffraction Facility supported by the MRSEC program of
the National Science Foundation at the Materials Research
Center at Northwestern University. We thank Dr. Jinsong Wu
for the TEM image and helpful discussions.
(36) Park, Y. M.; Daniel, J.; Heeney, M.; Salleo, A. Adv. Mater. 2011,
23, 971.
(37) DiBenedetto, S. A.; Frattarelli, D.; Ratner, M. A.; Facchetti, A.;
Marks, T. J. J. Am. Chem. Soc. 2008, 130, 7528.
(38) DiBenedetto, S. A.; Frattarelli, D. L.; Facchetti, A.; Ratner,
M. A.; Marks, T. J. J. Am. Chem. Soc. 2009, 131, 11080.
(39) Lee, B. H.; Ryu, M. K.; Choi, S. Y.; Lee, K. H.; Im, S.; Sung,
M. M. J. Am. Chem. Soc. 2007, 129, 16034.
’ REFERENCES
(1) Facchetti, A. Chem. Mater. 2011, 23, 733.
(2) Arias, A. C.; MacKenzie, J. D.; McCulloch, I.; Rivnay, J.; Salleo, A.
Chem. Rev. 2010, 110, 3.
(40) Tredgold, R. H. Nature 1991, 354, 120.
(41) Ulman, A. Chem. Rev. 1996, 96, 1533.
(42) Lee, H.; Kepley, L. J.; Hong, H. G.; Akhter, S.; Mallouk, T. E.
J. Phys. Chem. 1988, 92, 2597.
(43) Lee, H.; Kepley, L. J.; Hong, H. G.; Mallouk, T. E. J. Am. Chem.
Soc. 1988, 110, 618.
(3) Roberts, M. E.; Mannsfeld, S. C. B.; Queralto, N.; Reese, C.;
Locklin, J.; Knoll, W.; Bao, Z. N. Proc. Natl. Acad. Sci. U.S.A. 2008,
105, 12134.
(4) Braga, D.; Horowitz, G. Adv. Mater. 2009, 21, 1473.
10249
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250

Journal of the American Chemical Society
ARTICLE
(44) Katz, H. E.; Scheller, G.; Putvinski, T. M.; Schilling, M. L.;
Wilson, W. L.; Chidsey, C. E. D. Science 1991, 254, 1485.
(45) Katz, H. E.; Schilling, M. L. Chem. Mater. 1993, 5, 1162.
(46) Ha, Y. G.; Facchetti, A.; Marks, T. J. Chem. Mater. 2009,
21, 1173.
(81) Rivnay, J.; Jimison, L. H.; Northrup, J. E.; Toney, M. F.;
Noriega, R.; Lu, S. F.; Marks, T. J.; Facchetti, A.; Salleo, A. Nat. Mater.
2009, 8, 952.
(82) Fritz, S. E.; Kelley, T. W.; Frisbie, C. D. J. Phys. Chem. B 2005,
109, 10574.
(47) Lin, W. B.; Lin, W. P.; Wong, G. K.; Marks, T. J. J. Am. Chem.
Soc. 1996, 118, 8034.
(48) Zhu, P. W.; van der Boom, M. E.; Kang, H.; Evmenenko, G.;
Dutta, P.; Marks, T. J. Chem. Mater. 2002, 14, 4982.
(49) Yoon, M. H.; Facchetti, A.; Marks, T. J. Proc. Natl. Acad. Sci.
U.S.A. 2005, 102, 4678.
(83) Steudel, S.; De Vusser, S.; De Jonge, S.; Janssen, D.; Verlaak, S.;
Genoe, J.; Heremans, P. Appl. Phys. Lett. 2004, 85, 4400.
(84) Knipp, D.; Street, R. A.; Volkel, A. R. Appl. Phys. Lett. 2003,
82, 3907.
(85) Pernstich, K. P.; Haas, S.; Oberhoff, D.; Goldmann, C.;
Gundlach, D. J.; Batlogg, B.; Rashid, A. N.; Schitter, G. J. Appl. Phys.
2004, 96, 6431.
(50) Wang, L.; Yoon, M. H.; Lu, G.; Yang, Y.; Facchetti, A.; Marks,
T. J. Nat. Mater. 2006, 5, 893.
(51) Facchetti, A.; Abbotto, A.; Beverina, L.; van der Boom, M. E.; Dutta,
P.; Evmenenko, G.; Pagani, G. A.; Marks, T. J. Chem. Mater. 2003, 15, 1064.
(52) Kim, C.; Facchetti, A.; Marks, T. J. Science 2007, 318, 76.
(53) Parratt, L. G. Phys. Rev. 1954, 95, 359.
(54) Nelson, A. J. Appl. Crystallogr. 2006, 39, 273.
(55) Malik, A.; Lin, W.; Durbin, M. K.; Marks, T. J.; Dutta, P. J. Chem.
Phys. 1997, 107, 645.
(56) Fukuto, M.; Heilmann, R. K.; Pershan, P. S.; Yu, S. J. M.; Soto,
C. M.; Tirrell, D. A. J. Chem. Phys. 2003, 119, 6253.
(57) Lin, W. B.; Lee, T. L.; Lyman, P. F.; Lee, J. J.; Bedzyk, M. J.;
Marks, T. J. J. Am. Chem. Soc. 1997, 119, 2205.
(58) Ha, Y. G.; Jeong, S.; Wu, J. S.; Kim, M. G.; Dravid, V. P.;
Facchetti, A.; Marks, T. J. J. Am. Chem. Soc. 2010, 132, 17426.
(59) Byrne, P. D.; Facchetti, A.; Marks, T. J. Adv. Mater. 2008,
20, 2319.
(60) Kim, C.; Facchetti, A.; Marks, T. J. Adv. Mater. 2007, 19, 2561.
(61) Jeong, S.; Jeong, Y.; Moon, J. J. Phys. Chem. C 2008, 112, 11082.
(62) Kim, D.; Jeong, Y.; Song, K.; Park, S. K.; Cao, G. Z.; Moon, J.
Langmuir 2009, 25, 11149.
(63) Seo, S. J.; Hwang, Y. H.; Bae, B. S. Electrochem. Solid-State Lett.
2010, 13, H357.
(64) Chua, L. L.; Zaumseil, J.; Chang, J. F.; Ou, E. C. W.; Ho,
P. K. H.; Sirringhaus, H.; Friend, R. H. Nature 2005, 434, 194.
(65) Hulea, I. N.; Fratini, S.; Xie, H.; Mulder, C. L.; Iossad, N. N.;
Rastelli, G.; Ciuchi, S.; Morpurgo, A. F. Nat. Mater. 2006, 5, 982.
(66) Yoon, M. H.; Kim, C.; Facchetti, A.; Marks, T. J. J. Am. Chem.
Soc. 2006, 128, 12851.
(67) Jo, S. J.; Kim, C. S.; Kim, J. B.; Kim, J.; Lee, M. J.; Hwang, H. S.;
Baik, H. K.; Kim, Y. S. Appl. Phys. Lett. 2008, 93.
(68) Hwang, D. K.; Lee, K.; Kim, J. H.; Im, S.; Kim, C. S.; Baik, H. K.;
Park, J. H.; Kim, E. Appl. Phys. Lett. 2006, 88.
(69) Zirkl, M.; Haase, A.; Fian, A.; Schon, H.; Sommer, C.; Jakopic,
G.; Leising, G.; Stadlober, B.; Graz, I.; Gaar, N.; Schwodiauer, R.; Bauer-
Gogonea, S.; Bauer, S. Adv. Mater. 2007, 19, 2241.
(70) Whitesides, G. M.; Grzybowski, B. Science 2002, 295, 2418.
(71) Love, J. C.;Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides,
G. M. Chem. Rev. 2005, 105, 1103.
(72) Action, B. O.; Ting, G. G.; Shamberger, P. J.; Ohuchi, F. S.; Ma,
H.; Jen, A. K. Y. ACS Appl. Mater. Interfaces 2010, 2, 511.
(73) Fukuda, K.; Hamamoto, T.; Yokota, T.; Sekitani, T.;
Zschieschang, U.; Klauk, H.; Someya, T. Appl. Phys. Lett. 2009, 95.
(74) Brzoska, J. B.; Shahidzadeh, N.; Rondelez, F. Nature 1992,
360, 719.
(75) Tao, Y. T. J. Am. Chem. Soc. 1993, 115, 4350.
(76) Spori, D. M.; Venkataraman, N. V.; Tosatti, S. G. P.; Durmaz,
F.; Spencer, N. D.; Zurcher, S. Langmuir 2007, 23, 8053.
(77) Boulas, C.; Davidovits, J. V.; Rondelez, F.; Vuillaume, D. Phys.
Rev. Lett. 1996, 76, 4797.
(78) Virkar, A. A.; Mannsfeld, S.; Bao, Z. A.; Stingelin, N. Adv. Mater.
2010, 22, 3857.
(79) Yang, S. Y.; Shin, K.; Park, C. E. Adv. Funct. Mater. 2005,
15, 1806.
(80) Kobayashi, S.; Nishikawa, T.; Takenobu, T.; Mori, S.; Shimoda,
T.; Mitani, T.; Shimotani, H.; Yoshimoto, N.; Ogawa, S.; Iwasa, Y. Nat.
Mater. 2004, 3, 317.
10250
dx.doi.org/10.1021/ja202755x |J. Am. Chem. Soc. 2011, 133, 10239–10250