Printable cross-linked polymer blend dielectrics. Design strategies, synthesis, microstructures, and electrical properties, with organic field-effect transistors as testbeds

Article abstract of DOI:10.1021/ja801047g

We report here the synthesis and dielectric properties of optimized, cross-linked polymer blend (CPB) dielectrics for application in organic field-effect transistors (OFETs). Novel silane cross-linking reagents enable the synthesis of CPB films having excellent quality and tunable thickness (from 10 to ～500 nm), fabricated both by spin-coating and gravure-printing. Silane reagents of the formula X₃Si-R-SiX₃ (R = -C ₆H₁₂- and X = Cl, OAc, NMe₂, OMe, or R = -C₂H₄-O-C₂H₄- and X = OAc) exhibit tunable reactivity with hydroxyl-containing substrates. Dielectric films fabricated by blending X₃Si-R-SiX₃ with poly(4-vinyl)phenol (PVP) require very low-curing temperatures (～110°C) and adhere tenaciously to a variety of FET gate contact materials such as n ⁺-Si, ITO, and Al. The CPB dielectrics exhibit excellent insulating properties (leakage current densities of 10^-7 ～ 10^-8 A cm^-2 at 2.0 MV/cm) and tunable capacitance values (from 5 to ～350 nF cm^-2). CPB film quality is correlated with the PVP-cross-linking reagent reactivity. OFETs are fabricated with both p- and n-type organic semiconductors using the CPB dielectrics function at low operating voltages. The morphology and microstructure of representative semiconductor films grown on the CPB dielectrics is also investigated and is correlated with OFET device performance.

Full text of DOI:10.1021/ja801047g

Published on Web 05/03/2008
Printable Cross-Linked Polymer Blend Dielectrics. Design
Strategies, Synthesis, Microstructures, and Electrical
Properties, with Organic Field-Effect Transistors as Testbeds
Choongik Kim, Zhiming Wang, Hyuk-Jin Choi, Young-Geun Ha, Antonio Facchetti,*
and Tobin J. Marks*
Department of Chemistry and the Materials Research Center, Northwestern UniVersity,
2145 Sheridan Road, EVanston, Illinois 60208
Received February 14, 2008; E-mail: a-facchetti@northwestern.edu; t-marks@northwestern.edu
Abstract: We report here the synthesis and dielectric properties of optimized, cross-linked polymer blend
(CPB) dielectrics for application in organic field-effect transistors (OFETs). Novel silane cross-linking reagents
enable the synthesis of CPB films having excellent quality and tunable thickness (from 10 to ∼500 nm),
fabricated both by spin-coating and gravure-printing. Silane reagents of the formula X3Si-R-SiX3 (R )
-C6H12- and X ) Cl, OAc, NMe2, OMe, or R ) -C2H4-O-C2H4- and X ) OAc) exhibit tunable reactivity
with hydroxyl-containing substrates. Dielectric films fabricated by blending X3Si-R-SiX3 with poly(4-
vinyl)phenol (PVP) require very low-curing temperatures (∼110 °C) and adhere tenaciously to a variety of
FET gate contact materials such as n+-Si, ITO, and Al. The CPB dielectrics exhibit excellent insulating
properties (leakage current densities of 10-7 ∼ 10-8 A cm-2 at 2.0 MV/cm) and tunable capacitance values
(from 5 to ∼350 nF cm-2). CPB film quality is correlated with the PVP-cross-linking reagent reactivity.
OFETs are fabricated with both p- and n-type organic semiconductors using the CPB dielectrics function
at low operating voltages. The morphology and microstructure of representative semiconductor films grown
on the CPB dielectrics is also investigated and is correlated with OFET device performance.
Introduction
The development of polymeric gate dielectric materials has
been fundamental to the progress of organic electronic circuitry.1,2
Thus, emerging display and labeling technologies based on
organic field-effect transistors (OFETs), such as electronic paper/
posters and radiofrequency ID tags, require FET fabrication on
flexible plastic substrates over very large areas and via economi-
cal, high throughput processes.3,4 Such technologies will gain
wide acceptance only if electronic devices can be produced at
significantly lower cost than the current, capital-intensive
Figure 1. Schematic of the top-contact/bottom-gate OFET device geometry
employed in this study, and the structures of the polymers and silane cross-
linkers used for the fabrication of first-generation CPB gate dielectrics.12
(1) (a) Klauk, H. Organic Electronics: Materials, Manufacturing, and
Applications; Wiley-VCH: Weinheim, Germany, 2006. (b) Deman,
A. L.; Tardy, J. Mater. Sci. Eng., C 2006, 26, 421–426. (c) Klauk,
H.; Halik, M.; Zschieschang, U.; Schmid, G.; Radlik, W.; Weber, W.
J. Appl. Phys. 2002, 92, 5259–5263. (d) Sirringhaus, H.; Kawase, T.;
Friend, R. H.; Shimoda, T.; Inbasekaran, M.; Wu, W.; Woo, E. P.
Science 2000, 290, 2123–2126.
manufacturing technologies allow. Therefore, there have been
worldwide science-based efforts to create, understand, and
optimize FET component materials (semiconductor, dielectric,
and contactssFigure 1) which can be deposited via solution-
process methodologies such as spin-coating, casting, and
printing.5,6 Regarding organic semiconductors, many are readily
deposited from solution either directly or as soluble molecular/
polymeric precursors, which are then converted into the
insoluble semiconducting form.7 Regarding conductors, doped
conjugated polymers and nanoparticle-based conductive inks
allow solution fabrication of sufficiently low resistivity features
for source/drain and gate contact patterning.8 As far as gate
dielectrics are concerned, polymers are ideal candidates, and
(2) Veres, J.; Ogier, S.; Lloyd, G. Chem. Mater. 2004, 16, 4543–4555.
(3) (a) Jones, B. A.; Facchetti, A.; Marks, T. J.; Wasielewski, M. R. Chem.
Mater. 2007, 19, 2703–2705. (b) Ji, T.; Kathiresan, M.; Nair, S.; Jung,
S.; Natarajan, V.; Vishnubhatia, R. M. R.; Varadan, V. K. Proc. SPIE-
Int. Soc. Opt. Eng. 2007, 6528, 1–6. (c) Yoo, B.; Jung, T. O.; Basu,
D.; Dodabalapur, A.; Jones, B. A.; Facchetti, A.; Wasielewski, M. R.;
Marks, T. J. Appl. Phys. Lett. 2006, 88, 082104/1-3. (d) Subramanian,
V.; Chang, P. C.; Lee, J. B.; Molesa, S. E.; Volkman, S. K. IEEE
Trans. Compon. Packag. Technol. 2005, 28, 742–747. (e) Baude, P. F.;
Ender, D. A.; Haase, M. A.; Kelley, T. W.; Muyres, D. V.; Theiss,
S. D. Appl. Phys. Lett. 2003, 82, 3964–3966. (f) Drury, C. J.; Mutsaers,
C. M. J.; Hart, C. M.; Matters, M.; de Leeuw, D. M. Appl. Phys. Lett.
1998, 73, 108–110. (g) Gamota, D. R.; Brazis, P.; Kalyanasundaram,
X.; Zhang, J. Printed Organic and Molecular Electronics; Kluwer
Academic Publishers: New York, 2004.
(4) Singh, T. B.; Sariciftci, N. S. Annu. ReV. Mater. Res. 2006, 36, 199–
230.
9
10.1021/ja801047g CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 6867–6878 6867

A R T I C L E S
Kim et al.
most of the polymer dielectrics employed to date for OFETs
are commercially available, exhibiting a large range of dielectric
constants and acceptable solubility in a variety of solvents.9,10
A typical OFET structure is shown in Figure 1. In this device,
the conductance of the source-drain channel region is modulated
by the source-gate electric field (EG). When the device is in the
off-state (EG ) 0) the channel conductance is very low (typically
<10-12 S), whereas in the on-state (EG * 0) a sharp increase
in conductance is observed (>10-6 S), with output currents (in
saturation) defined as in11
current on/off ratio (Ion/Ioff). The gate insulator affects device
performance in many ways, principally permitting the creation
of the gate field and establishing a two-dimensional channel
charge sheet. The accumulated charge carriers transit from the
source to the drain electrodes in an area very close to the
dielectric-semiconductor interface upon application of a source-
drain bias. Therefore, the nature of the semiconductor-dielectric
interface, hence the dielectric surface morphology prior to
semiconductor deposition, greatly affects how the accumulated
charges move in the semiconductor and, therefore, the field-
effect mobility.13 Furthermore, the dielectric morphology and
surface energy tuning (e.g., surface treatment via a self-
assembled monolayer) have been shown to significantly modify
the growth, morphology, and microstructure of the overlying
vapor/solution-deposited semiconductorsfactors all affecting µ
and Ion/Ioff.14 The gate insulator dielectric properties also affect
the proximate semiconductor density of states distribution for
both amorphous and single-crystal semiconductors.15 Also
critical for long-term device stability is the extent to which the
dielectric surface traps charge carriers via reactive chemical
groups, and the density of interfacial states and charged
species.16
W
2L
2
IDS
)
µC V - V
)
T
(1)
(
i
G
where W/L is the channel width/length, Ci is the dielectric
capacitance per unit area, VG is the source-gate voltage, and VT
is the threshold voltage. Among the key device parameters are
the semiconductor field-effect mobility (µ) and the device
(5) (a) Herlogsson, L.; Crispin, X.; Robinson, N. D.; Sandberg, M.; Hagel,
O.-J.; Gustafsson, G.; Berggren, M. AdV. Mater. 2007, 19, 97–101.
(b) Yildirim, F. A.; Ucurum, C.; Schliewe, R. R.; Bauhofer, W.;
Meixner, R. M.; Goebel, H.; Krautschneider, W. Appl. Phys. Lett. 2007,
90, 083501/1-3. (c) Itoh, E.; Miyairi, K. Thin Solid Films 2006, 499,
95–103. (d) Kim, Y. M.; Lim, E.; Kang, I.-N.; Jung, B.-J.; Lee, J.;
Koo, B. W.; Do, L.-M.; Shim, H.-K. Macromolecules 2006, 39, 4081–
4085. (e) Yang, S. Y.; Shin, K.; Park, C. E. AdV. Funct. Mater. 2005,
15, 1806–1814. (f) Naber, R. C. G.; Tanase, C.; Blom, P. W. M.;
Gelinck, G. H.; Marsman, A. W.; Touwslager, F. J.; Setayesh, S.; de
Leeuw, D. M. Nat. Mater. 2005, 4, 243–248. (g) Liu, Y.; Varahramyan,
K.; Cui, T. Macromol. Rapid Commun. 2005, 26, 1955–1959. (h)
Granstrom, J.; Katz, H. E. J. Mater. Res. 2004, 19, 3540–3546. (i)
Shtein, M.; Peumans, P.; Benziger, J. B.; Forrest, S. R. AdV. Mater.
2004, 16, 1615–1620. (j) Knobloch, A.; Manuelli, A.; Bernds, A.;
Clemens, W. J. Appl. Phys. 2004, 96, 2286–2291.
There are three major limitations to current-generation
polymer gate dielectric-based OFETs. The first is that, with few
exceptions,12,17 they function at relatively large operating
voltages, reflecting the intrinsically low (compared to crystalline
(13) (a) Park, Y. D.; Lim, J. A.; Lee, H. S.; Cho, K. Mater. Today 2007,
10, 46–54. (b) Surin, M.; Leclere, Ph.; Lazzaroni, R.; Yuen, J. D.;
Wang, G.; Moses, D.; Heeger, A. J.; Cho, S.; Lee, K. J. Appl. Phys.
2006, 100, 033712/1-6. (c) Lee, C. A.; Park, D. W.; Jin, S. H.; Park,
I. H.; Lee, J. D.; Park, B.-G. Appl. Phys. Lett. 2006, 88, 252102/1-3.
(d) Chua, L.-L.; Zaumsell, J.; Chang, J.-F.; Ou, E. C.-W.; Ho, P. K.-
H.; Sirringhaus, H.; Friend, R. H. Nature 2005, 434, 194–199. (e)
Panzer, M. J.; Frisbie, C. D. J. Am. Chem. Soc. 2005, 127, 6960–
6961. (f) Schroeder, R.; Majewski, L. A.; Grell, M. AdV. Mater. 2005,
17, 1535–1539. (g) Kobayashi, S.; Nishikawa, T.; Takenobu, T.; Mori,
S.; Shimoda, T.; Mitani, T.; Shimotani, H.; Yoshimono, N.; Ogawa,
S.; Iwasa, Y. Nat. Mater. 2004, 3, 317–322. (h) Pernstich, K. P.; Haas,
S.; Oberhoff, D.; Goldmann, C.; Gundlach, D. J.; Batlogg, B.; Rashid,
A. N.; Schitter, G. J. J. Appl. Phys. 2004, 96, 6431–6438. (i)
Dimitrakopoulos, C. D.; Purushothaman, S.; Kymissis, J.; Callegari,
A.; Shaw, J. M. Science 1999, 283, 822–824.
(6) Bao, Z.; Rogers, J. A.; Katz, H. E. J. Mater. Chem. 1999, 9, 1895–
1904.
(7) (a) Facchetti, A. Mater. Today 2007, 10, 28–37. (b) Payne, M. M.;
Parkin, S. R.; Anthony, J. E.; Kuo, C.-C.; Jackson, T. N. J. Am. Chem.
Soc. 2005, 127, 4986–4987. (c) Afzali, A.; Kagan, C. R.; Traub, G. P.
Synth. Met. 2005, 155, 490–494. (d) Afzali, A.; Dimitrakopoulos,
C. D.; Breen, T. L. J. Am. Chem. Soc. 2002, 124, 8812–8813. (e)
Herwig, P. T.; Mu¨llen, K. AdV. Mater. 1999, 11, 480–483. (f) Brown,
A. R.; Pomp, A.; de Leeuw, D. M.; Klaassen, D. B. M.; Havinga,
E. E.; Herwig, P.; Mu¨llen, K. J. Appl. Phys. 1996, 79, 2136–2138.
(8) (a) Gamerith, S.; Klug, A.; Scheibrer, H.; Scherf, U.; Moderegger,
E.; List, E. J. W. AdV. Funct. Mater. 2007, 17, 3111–3118. (b) Kim,
D.; Jeong, S.; Moon, J.; Han, S.; Chung, J. Appl. Phys. Lett. 2007,
91, 071114/1-3. (c) Ko, S. H.; Pan, H.; Grigoropoulos, C. P.;
Luscombe, C. K.; Frechet, J. M. J.; Poulikakos, D. Appl. Phys. Lett.
2007, 90, 141103/1-3. (d) Noguchi, Y.; Sekitani, T.; Someya, T. Appl.
Phys. Lett. 2006, 89, 253507/1-3. (e) Stutzmann, N.; Friend, R. H.;
Sirringhaus, H. Science 2003, 299, 1881–1884. (f) Pron, A.; Rannou,
P. Prog. Polym. Sci. 2002, 27, 135–190.
(14) (a) Yoon, M.-H.; Kim, C.; Facchetti, A.; Marks, T. J. J. Am. Chem.
Soc. 2006, 128, 12851–12869. (b) Hamadani, B. H.; Corley, D. A.;
Ciszek, J. W.; Tour, J. M.; Natelson, D. Nano Lett. 2006, 6, 1303–
1306. (c) Lee, H. S.; Kim, D. H.; Cho, J. H.; Park, Y. D.; Kim, J. S.;
Cho, K. AdV. Funct. Mater. 2006, 16, 1859–1864. (d) Tsukagoshi,
K.; Shigeto, K.; Yagi, I.; Aoyagi, Y. Appl. Phys. Lett. 2006, 89,
113507/1-3. (e) DeLongchamp, D. M.; Sambasivan, S.; Fischer,
D. A.; Lin, E. K.; Chang, P.; Murphy, A. R.; Frechet, J. M. J.;
Subramanian, V. AdV. Mater. 2005, 17, 2340–2344. (f) Salleo, A.;
Chabinyc, M. L.; Yang, M. S.; Street, R. A. Appl. Phys. Lett. 2002,
81, 4383–4385. (g) Lin, Y. Y.; Gundlach, D. J.; Nelson, S. F.; Jackson,
T. N. IEEE Electron DeVice Lett. 1997, 18, 606–608.
(9) Facchetti, A.; Yoon, M.-H.; Marks, T. J. AdV. Mater. 2005, 17, 1705–
1725.
(10) (a) Kaake, L. G.; Zou, Y.; Panzer, M. J.; Frisbie, C. D.; Zhu, X.-Y.
J. Am. Chem. Soc. 2007, 129, 7824–7830. (b) Noh, S. H.; Choi, W.;
Oh, M. S.; Hwang, D. K.; Lee, K.; Im, S.; Jang, S.; Kim, E. Appl.
Phys. Lett. 2007, 90, 253504/1-3. (c) Bae, J.-H.; Kim, J.; Kim, W.-
H.; Lee, S.-D. Jpn. J. Appl. Phys. 2007, 46, 385–389. (d) Zhao, Y.;
Dong, G.; Wang, L.; Qiu, Y. Appl. Phys. Lett. 2007, 90, 252110/1-
3. (e) Ghim, J.; Baeg, K.-J.; Noh, Y.-Y.; Kang, S.-J.; Jo, J.; Kim,
D.-Y.; Cho, S.; Yuen, J.; Lee, K.; Heeger, A. J. Appl. Phys. Lett. 2006,
89, 202516/1-3. (f) Lee, K.; Kim, J. H.; Im, S. Appl. Phys. Lett. 2006,
88, 023504/1-3. (g) Unni, K. N. N.; Dabos-Seignon, S.; Nunzi, J.-
M. J. Mater. Sci. 2006, 41, 317–322. (h) Sung, J. H.; Park, S. J.; Park,
J. H.; Choi, H. J.; Choi, J. S. Synth. Met. 2006, 156, 861–864. (i) Lee,
J.; Hwang, D. K.; Choi, J.-M.; Lee, K.; Kim, J. H.; Im, S.; Park, J. H.;
Kim, E. Appl. Phys. Lett. 2005, 87, 023504/1-3. (j) Park, S. Y.; Park,
M.; Lee, H. H. Appl. Phys. Lett. 2004, 85, 2283–2285. (k) Gelinck,
G. H.; Geuns, T. C. T.; de Leeuw, D. M. Appl. Phys. Lett. 2000, 77,
1487–1489.
(15) (a) Wang, A.; Kymissis, I.; Bulovic, V.; Ibitayo, A. IEEE Trans.
Electron DeVices 2006, 53, 9–13. (b) Lang, D. V.; Chi, X.; Siegrist,
T.; Sergent, A. M.; Ramirez, A. P. Phys. ReV. Lett. 2004, 93, 086802/
1-4.
(16) (a) Goldmann, C.; Krellner, C.; Pernstich, K. P.; Haas, S.; Gundlach,
D. J.; Batlogg, B. J. Appl. Phys. 2006, 99, 034507/1-8. (b) Gomes,
H. L.; Stallinga, P.; Colle, M.; Biscarini, F.; de Leeuw, D. M. J. Non-
Cryst. Solids 2006, 352, 1761–1764. (c) Gu, G.; Kane, M. G.; Doty,
J. E.; Firester, A. H. Appl. Phys. Lett. 2005, 87, 243512/1-3.
(17) (a) 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, 243513/1-3. (b) Yang,
S. Y.; Kim, S. H.; Shin, K.; Jeon, H.; Park, C. E Appl. Phys. Lett.
2006, 88, 173507/1-3. (c) Jang, Y.; Kim, D. H.; Park, Y. D.; Cho,
J. H.; Hwang, M.; Cho, K. Appl. Phys. Lett. 2006, 88, 072101/1-3.
(d) Panzer, M. J.; Newman, C. R.; Frisbie, C. D. Appl. Phys. Lett.
2005, 86, 103503/1-3.
(11) Sze, S. M. Semiconductor DeVices: Physics and Technology; 2nd ed.;
Wiley: New York, 1981.
(12) Yoon, M.-H.; Yan, H.; Facchetti, A.; Marks, T. J. J. Am. Chem. Soc.
2005, 127, 10388–10395.
9
6868 J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008

Printable Cross-Linked Polymer Blend Dielectrics
A R T I C L E S
inorganic semiconductors) field-effect mobilities.18 Second, very
few polymeric dielectrics perform optimally with a broad variety
of both hole- and electron-transporting organic semiconductors,
limiting organic complementary circuit applications. Third, it
is very difficult with the same polymer formulation to realize
nanoscopically thin films exhibiting large capacitance values
but having minimal gate leakage current at useful gate fields.
From eq 1, an attractive means to address the first problem,
while maintaining all the other parameters constant, would be
to substantially increase the gate Ci, which would allow
achieving the same current but at significantly lower VG values.
Since
k
Ci ) ε
(2)
0d
Figure 2. Structures of the organic semiconductor (left) and of the CPB
precursors (polymer and silane cross-linkers, right) employed in this study.
where k is the dielectric constant, ε0 is the vacuum permittivity,
and d is the insulator thickness. Ci is increased as k increases
and/or d decreases. However, the k value of most insulating
polymers is confined within a relatively narrow range (ca. 2-6)
and, equally important, the films must be very thick (usually
∼1 µm) to avoid significant leakage currents to the gate
electrode.9,19 To reduce leakage currents for thinner films, cross-
linked polymer dielectrics such as melamine-poly(4-vinyl)phe-
nol (PVP) and cross-linked BCB (benzocyclobutene) have been
introduced,20 which allow reducing the dielectric thickness.
However, films of these materials require high annealing
temperatures (>150-200 °C) and Ci values are typically <20
reactivity with respect to hydroxyl-containing dielectric poly-
mers (Figure 2). For the first time, an in situ kinetic study
comparing the reactivity of chloro-, acetoxy-, dialkylamino-,
and methoxy-silanes with a hydroxylated reactant is reported.
This study thereby affords a correlation of cross-linker-polymer
reactivity patterns with the corresponding CPB dielectric
properties. With the new silane cross-linkers, we demonstrate
that optimization of cross-linking conditions, especially by using
moderately reactive cross-linking reagents, appropriate solvents,
and polymer/cross-linker molar ratio, affords very low leakage
and Very smooth dielectric films. These new CPB formulations
enable tunable thickness and capacitances and can be fabricated
on conventional Si substrates as well as on flexible Al/PEN and
ITO/Mylar substrates via spin-coating and gravure printing.
OFETs fabricated with these new CPB dielectric films exhibit
good performance for both p-(pentacene) and n-type (DFHCO-
4T21) organic semiconductors (Figure 2). Finally, the hydrophilic
properties of the polymer and siloxane matrices in the CPB films
are correlated with humidity-dependent dielectric properties and
the resulting OFET device response.
nF cm-2
.
Recently, we demonstrated that blending appropriate π-elec-
tron polymers, for example, PVP and PS, with chlorosilane
cross-linking reagents affords robust, adherent, pinhole-free,
high-capacitance, low-leakage, ultrathin (10-20 nm) gate
dielectric materials (Figure 1).12 However, owing to the high
reactivity of the chlorosilane cross-linking reagents and unop-
timized cross-linking reactions, the resulting dielectric materials
prepared by spin-coating and post-annealing exhibit less than
optimum surface morphologies (rms roughness > 2 nm),
compromising OFET performance, especially the field-effect
charge carrier mobility. Furthermore, to enable deposition of
the CPB films via roll-to-roll processes, hence by printing
methodologies, it is of fundamental importance to control the
kinetics of cross-linking to avoid premature cross-linked matrix
formation in solution before film deposition. Note that with
current printing technologies, relatively thick (a few µms)
dielectric films can be deposited for many applications.3g
In this contribution, we report on the design, synthesis, and
implementation of new silane cross-linkers having tuned
Results
Strategy for Cross-Linking Reagent Design. The CPB materi-
als reported here consist of poly(4-vinyl)phenol (PVP) chains
covalently bound in cross-linked matrices (Figure 2). As in the
case of other cross-linked polymers employed for diverse
applications,12,20 this methodology enhances structural robust-
ness and, for organic electronic applications, electrical insulating
properties. The present cross-linking strategy employs the
coupling reaction of selected bifunctional organosilane reagents
with OH-functionalized molecules/polymers and/or with H2O
to produce robust siloxane networks (Scheme 1). We previously
reported that when chlorosilanes are used in combination with
PVP (or polystyrene), cross-linked matrices are produced in air
immediately after film deposition and curing. Moreover, the film
structural integrity and connectivity of these materials are
evident by the complete insolubility of the product in the mother
solution (Figure 2).12 The curing temperature (∼110 °C) is fully
compatible with the common varieties of plastic substrates
employed in organic electronics.1a,3g
(18) Dimitrakopoulos, C. D.; Malenfant, P. R. L. AdV. Mater. 2002, 14,
99–117.
(19) (a) Singh, T. B.; Meghdadi, F.; Gu¨nes, S.; Marjanovic, N.; Horowitz,
G.; Lang, P.; Bauer, S.; Sariciftci, N. S. AdV. Mater. 2005, 17, 2315–
2320. (b) Nunes, G., Jr.; Zane, S. G.; Meth, J. S. J. Appl. Phys. 2005,
98, 104503/1-6.
(20) (a) Lim, S. C.; Kim, S. H.; Koo, J. B.; Lee, J. H.; Ku, C. H.; Yang,
Y. S.; Zyung, T. Appl. Phys. Lett. 2007, 90, 173512/1-3. (b) Lee,
S. H.; Choo, D. J.; Han, S. H.; Kim, J. H.; Son, Y. R.; Jang, J. Appl.
Phys. Lett. 2007, 90, 033502/1-3. (c) Kim, S. H.; Yang, S. Y.; Shin,
K.; Jeon, H.; Lee, J. W.; Hong, K. P.; Park, C. E. Appl. Phys. Lett.
2006, 89, 183516/1-3. (d) Marjanovic, N.; Singh, T. B.; Dennler,
G.; Gu¨nes, S.; Neugebauer, H.; Sariciftci, N. S.; Schwo¨diauer, R.;
Bauer, S. Org. Electron. 2006, 7, 188–194. (e) Jang, Y.; Kim, D. H.;
Park, Y. D.; Cho, J. H.; Hwang, M.; Cho, K. Appl. Phys. Lett. 2005,
87, 152105/1-3. (f) Singh, T. B.; Marjanovic, N.; Stadler, P.; Auinger,
M.; Matt, G. J.; Gu¨nes, S.; Sariciftci, N. S.; Schwo¨diauer, R.; Bauer,
S. J. Appl. Phys. 2005, 97, 083714/1-5. (g) Chua, L.-L.; Ho, P. K. H.;
Sirringhaus, H.; Friend, R. H. Appl. Phys. Lett. 2004, 84, 3400–3402.
Previous CPB dielectrics based on chlorosilane cross-linkers
afford relatively rough dielectric film surfaces (rms roughness
>2 nm for 10-20 nm thick films) due to the great reactivity of
(21) Yoon, M.-H.; DiBenedetto, S. A.; Facchetti, A.; Marks, T. J. J. Am.
Chem. Soc. 2005, 127, 1348–1349.
9
J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008 6869

A R T I C L E S
Kim et al.
Scheme 2. Synthesis of the New CPB Crosslinking Reagents
(C6OAc, C6NMe2, EGOAc) and Model Substrates Used for the
Kinetic Study (C6SiOAc, C6SiNMe2, and C6SiOMe)
Scheme 1. Schematic of the Crosslinking Processes for the
Various Organosilane Crosslinking Reagents Employed in This
Study
the cross-linking reagents.12 Furthermore, for the same reason,
chlorosilane-based dielectric formulations cannot be readily
printed (vide infra). To optimize cross-linking reaction condi-
tions, leading to robust, printable films and smoother dielectric
surface morphologies, as well as to assess the influence of
different CPB cross-linking formulations, the combination of
PVP with a variety of R-ω-hexyl cross-linking reagents, differing
in the cross-linkable silane end groups and core structures, was
explored, including 1,6-bis(trichlorosilyl)hexane (C6Cl; as a
reference), 1,6-bis(triacetoxysilyl)hexane (C6OAc), 1,6-bis[tri-
(dimethylamino)silyl]hexane (C6NMe2), and 1,6-bis(trimethoxy-
silyl)hexane (C6OMe; Figure 2). The R,ω-hexyl (C6) linker was
chosen because it was previously found to afford the smoothest
film surface morphology when used as a chlorosilane-based
cross-linker.12 The dioxyethylene-containing core in the EGOAc
cross-linker was selected because of the improved chemical
compatibility with PVP, leading to more complete cross-linking
and smoother dielectric surfaces (vide infra).
Scheme 3. Model Reactions of Functional Silanes (C6SiMe2Cl,
C6SiMe2OAc, C6SiMe2NMe2, and C6SiMe2OMe) with
4-Isopropylphenol (PhOH)
Cross-Linker Synthesis. The reagents 1,6-bis(triacetoxysilyl)-
hexane (C6OAc), bis(3-triacetoxysilylpropyl ethylene ether)
(EGOAc), 1,6-bis(tri(dimethylamino)silyl)hexane (C6NMe2),
dimethylhexylacetoxysilane (C6SiMe2OAc), dimethylhexyl(di-
methylamino)silane (C6SiMe2NMe2) and dimethylhexylmetho-
xysilane (C6SiMe2OMe) were synthesized as shown in Scheme
2. C6OAc and C6NMe2 were obtained in almost quantitative
yields by reaction of 1,6-bis(trichlorosilyl)hexane with acetic
anhydride and dimethylamine, respectively. Similarly, C6Si-
Me2OAc, C6SiMe2NMe2, and C6SiMe2OMe (model reagents
employed for the kinetic study, vide infra) were synthesized
by reaction of commercially available dimethylhexylchlorosilane
(C6SiMe2Cl) with acetic anhydride, dimethylamine, or anhy-
drous methanol, respectively. EGOAc was synthesized in three
steps: (1) allylation of 2-allyloxy-ethanol in the presence of allyl
bromide and sodium hydride (or anhydrous sodium carbonate)
to form diallylethylene ether; (2) hydrosilylation of diallyleth-
ylene ether with trichlorosilane to afford [bis(3-trichlorosilyl)-
propyl] ethylene ether; (3) functional group conversion of [bis(3-
trichlorosilyl)propyl] ethylene ether with acetic anhydride to
yield EGOAc in 60% overall yield. The chemical structures and
purities of the new silane cross-linking reagents were verified
proposed. The presence of catalysts may lead to more complex
reaction kinetics and mechanisms, which depend on the nu-
cleophilicity of the catalyst and the reaction stereochemistry.22
The surface morphologies of the cross-linked dielectric films
depend strongly on the PVP coupling rate with the chosen silane
cross-linker. Therefore, the reactivities of the silane cross-linker
leaving groups (chloro-, acetoxy-, dimethylamino-, and meth-
oxy) were quantified using model reactions. To simplify the
kinetic study of the coupling reactions, monofunctional silanes,
including dimethylhexylchlorosilane (C6SiMe2Cl), dimethyl-
hexylacetoxysilane (C6SiMe2OAc), dimethylhexyl(dimethylami-
no)silane (C6SiMe2NMe2), and dimethylhexylmethoxysilane
(C6SiMe2OMe) were used to model the cross-linker reagents,
and 4-isopropylphenol (PhOH) was chosen as a substrate to
model PVP (Scheme 3). In contrast to the kinetic studies at
room temperature in the presence of condensation catalysts
reported by others,22 reactions here were carried out at 110 °C
without added catalyst, similar to the conditions employed for
the polymer-cross-linker reactions leading to the CPB dielectric
films.
1
by H and 29Si NMR as well as by elemental analysis.
Kinetic Study. The kinetics of nucleophilic substitution
reactions of organosilicon compounds have been reported in
previous literature. In most studies, halosilanes and water (or
alcohols) were chosen as substrates and nucleophiles, respec-
tively. In absence of catalysts, a pseudo-first-order chlorosilane
hydrolysis (alcoholysis) reaction was studied using conventional
techniques such as sampling and titration. Second-order rate
coefficients (k ≈ 10-4 L mol-1 s-1 for chlorotriisopropylsilane
in dioxane) were determined and an SN2-type mechanism
The product of the reaction is dimethylhexyl-4-isopropylphe-
noxysilane (C6SiMe2OPh), independent of the silane reagent.
(22) (a) Sommer, L. H. Stereochemistry, Mechanism and Silicon; McGraw-
Hill, Inc.: New York, 1965. (b) Rubinsztajn, S.; Cypryk, M.;
Chojnowski, J. J. Organomet. Chem. 1989, 367, 27–37. (c) Corrju,
R. J. P.; Dabosi, G.; Martineau, M. J. Chem. Soc., Chem. Commun.
1977, 649–650.
9
6870 J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008

Printable Cross-Linked Polymer Blend Dielectrics
A R T I C L E S
Table 1. 1H NMR and Kinetic Data for Model Reactions of Functional Silanes with 4-Isopropylphenol (PhOH)
Compound
1H NMR (CDCl3)
k (L mol-1 s-1
)
C6SiMe2NMe2
CH3CH2CH2CH2CH2CH2Si(CH3)2(N(CH3)2)
too fast to measure
2.42 (s, 6H), 1.32 (m, 8H), 0.88 (t, 3H), 0.58 (t, 2H), 0.08 (s, 6H)
CH3CH2CH2CH2CH2CH2Si(CH3)2Cl
1.42-1.15 (m, 8H), 0.85 (t, 3H), 0.78 (t, 2H), 0.38 (s, 6H)
CH3CH2CH2CH2CH2CH2Si(CH3)2(OCOCH3)
2.03 (s, 3H), 1.45-1.18 (m, 8H), 0.86 (t, 3H), 0.72 (t, 2H), 0.22 (s, 6H)
CH3CH2CH2CH2CH2CH2Si(CH3)2(OCH3)
3.55 (s, 3H), 1.32 (m, 8H), 0.90 (t, 3H), 0.52 (t, 2H), 0.05 (s, 6H)
CH3CH2CH2CH2CH2CH2Si(CH3)2(OPhCH(CH3)2)
7.09 (d, 2H), 6.77 (d, 2H), 2.85 (m, 1H), 1.40-1.18 (m, 14H), 0.86 (t,
3H), 0.72 (t, 2H), 0.23 (s, 6H)
C6SiMe2Cl
8.78 × 10-5
4.37 × 10-6
no reaction
C6SiMe2OAc
C6SiMe2OMe
C6SiMe2OPh
PhOH
HOPhCH(CH3)2
7.10 (d, 2H), 6.85 (d, 2H), 4.85 (s, 1H), 2.85 (m, 1H), 1.20 (d, 6H)
The chemical shift of the Si-CH3 groups is sensitive to the
leaving group (Table 1). For example, as the reaction proceeds,
the resonance intensity at δ 0.38 ppm assigned to the C6SiMe2Cl
methyl groups decreases, while the product C6SiMe2OPh methyl
signal located at δ 0.23 ppm increases in intensity (Figure 3).
Hence, based on the Si-CH3 group peak intensities, the
concentration of monofunctional silane reagents was monitored
over time (2 h) by in situ 1H NMR. The model reaction kinetics
obey the empirical rate law: Rp ) k[PhOH][C6SiMe2X] over a
wide range of concentrations where Rp is reaction rate and k is
rate constant (Figure 4). As shown in Table 1, the reactivity of
the silane functional group decreases in the order NMe2 . Cl
> OAc . OMe. The rate constant for C6SiMe2Cl (8.78 × 10-5
L mol-1 s-1) is comparable to (slightly lower than) the
alcoholysis rate constant of chlorotriisopropylsilane (∼1.5 ×
10-4 L mol-1 s-1), but less than that of chlorotriisopropylsilane
(∼6 × 10-4 L mol-1 s-1). As will be shown later, these data
are in excellent accord with the observed CPB surface mor-
phological characteristics, film robustness, and dielectric char-
acteristics of the films fabricated with the various silane cross-
linkers (vide infra).
CPB Film Fabrication and Characterization. To enable
deposition of CPB films via printing, it is important to control
and optimize the cross-linking conditions to produce robust films
with smooth surface morphologies. Moreover, acceptable OFET
response parameters must be achieved using concentrated
polymer/cross-linker solutions having appropriate rheologies for
printing. In previous studies, the relatively thick CPB films
fabricated by gravure printing and spin-coating12 using highly
reactive chlorosilane-based cross-linking reagents (vide infra)
did not afford optimal microstructural and dielectric properties.
Therefore, spin-coated ultrathin (<20 nm) CPB films were first
investigated in the present study to determine how different
cross-linking reagent reactivities affect CPB film dielectric
properties. Similar optimization was then performed for thicker
(50-300 nm) spin-coated CPB films to understand how/if
thicker CPB film properties differ from those of the correspond-
ing ultrathin films. Finally, the cross-linking reagents affording
optimum dielectric properties for spin-coated thick CPB films
were selected for gravure-printing experiments. For comparison
purposes, all leakage current density and capacitance measure-
ments are carried out under vacuum and selected samples then
are tested in various atmospheres.
a. Spin-Coated Ultrathin (<20 nm) CPB Films. When em-
ploying a new cross-linker, it is first necessary to reoptimize
film deposition conditions by varying parameters such as solvent
type and polymer/cross-linker stoichiometric ratio. The goal is
to achieve, for a specific film thickness range, optimal film
robustness and a smooth dielectric surface. Since a smooth
dielectric-semiconductor interface is a prerequisite for efficient
charge transport in the FET channel,23 the visual appearance
of the film combined with AFM data provide important
information for initial materials screening and evaluation. Table
2 summarizes an example describing the optimization of ∼15
nm-thick spin-coated CPB films fabricated with PVP-C6OAc,
by varying the deposition solvent and polymer/cross-linker ratio.
Note that a similar procedure was performed for each cross-
linker. Independent of the polymer/cross-linker ratio, EtOAc
was found to be the optimum solvent for depositing CPB films
prepared from PVP-C6OAc, affording very smooth surfaces (rms
Figure 3. 1H NMR spectra of a C6SiMe2Cl + PhOH mixture recorded at
different reaction times (t ) 1, 15, and 120 min, from bottom to top) in
C2D2Cl4 at 110 °C.
(23) (a) Fritz, S. E.; Kelley, T. W.; Frisbie, C. D. J. Phys. Chem. B 2005,
109, 10574–10577. (b) Knipp, D.; Street, R. A.; Volkel, A. R. Appl.
Phys. Lett. 2003, 82, 3907–3909.
Figure 4. Typical kinetic plot of ln([C6SiMe2X]/[C6SiMe2X]0) versus
reaction time with PhOH.
9
J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008 6871

A R T I C L E S
Kim et al.
Table 2. PVP-C6OAc Film Deposition Conditionsa
It is found that optimized polymer/cross-linker weight ratios
vary somewhat with the cross-linker. However, for all cross-
linkers, the estimated optimum PVP hydroxyl group (-OH):
cross-linker molar ratios are ∼3:1. THF, dioxane, and EtOAc
were used for solvent optimization, with EtOAc affording the
smoothest optimized CPB dielectric films for all cross-linkers
except C6Cl. Even though all the optimized CPB dielectric films
afford rather smooth morphologies (rms roughness <1 nm)
when an optimized solvent and polymer/cross-linker ratio is
employed, there are still significant rms roughness variations.
C6NMe2 (F ) 0.7-0.8 nm) and C6Cl (F ) 0.5-0.6 nm), with
the most reactive silane groups afford rougher morphologies
compared to moderately reactive C6OAc (F ) 0.3-0.4 nm) and
EGOAc (F ) 0.2-0.3 nm). These CPB morphological variations
with the cross-linker type correlate with the differing OFET
performance metrics observed for both p- and n-type semicon-
ductors (vide infra). Similarly, the different cross-linkers afford
CPB films with differing electrical properties. Figure 6 presents
leakage current density data for various optimized CPB dielectric
films. As shown, the C6OAc- and EGOAc-based CPB films
afford slightly lower leakage current densities (J ) 2-20 ×
10-8 A/cm2 at 2.0 MV/cm) compared to films prepared with
C6Cl and C6NMe2 cross-linkers (J ) 8-80 × 10-8 A/cm2 at
2.0 MV/cm). As expected, C6OMe affords far more leaky films
(J > 1 × 10-5 A/cm2 at 2.0 MV/cm). All of the spin-coated
thin CPB films except the ones fabricated with C6OMe afford
breakdown fields between 3-6 MV/cm, independent of the
cross-linker (Figure 6B). The areal capacitance values for spin-
coated thin and thick CPB films (vide infra) were measured as
a function of frequency (Ci-f, 1-103 kHz) and voltage (Ci-V)
on metal-insulator-semiconductor (MIS, M ) Au, S ) n+-
Si) structures. Capacitance values of CPB films fabricated using
C6OMe as the cross-linker cannot be measured because of the
large leakage currents. Representative Ci-f plots are shown in
Figure 6C, with the data (Table 3) demonstrating that spin-
coated thin CPB films afford large areal capacitance values
(300-350 nFcm-2 at 10 kHz). The formal dielectric constants
calculated from the capacitance values using eq 2, are 5.1-5.7
(not shown), which are reasonable considering that the reported
dielectric constants of PVP and cross-linked PVP dielectric films
are 3.6-6.1.12,20,25
spin-coating solvent (2 mL)
PVP/C6OAc ratio (mg/mg)
D
F
Cib
THF
4:8
4:6
4:4
4:0
4:8
4:6
4:4
4:0
4:8
4:6
4:4
4:0
18
17
14
22
15
13
10
21
17
13
12
20
3.0∼3.5
2.0∼2.5
2.0∼2.5
∼0.6
260
270
330
195
310
340
365
200
270
345
350
220
dioxane
EtOAc
4.0∼4.5
1.0∼1.5
1.0∼1.5
∼0.6
0.4∼0.5
0.3∼0.4
0.3∼0.4
0.3∼0.4
a Solvent, polymer/crosslinker ratio, film thickness (D, nm), RMS
roughness (F, nm), and areal capacitance values (Ci, nF cm-2).
b Measured in an MIS structure.
Figure 5. AFM images of PVP-C6OAc (4:6)-based CPB dielectric films
deposited from (A) THF, (B) dioxane, and (C) EtOAc.
roughness ) 0.3-0.4 nm). Figure 5 shows tapping mode AFM
images of spin-coated PVP-C6OAc films deposited from THF,
dioxane, and EtOAc. Clearly, films fabricated from THF and
dioxane afford relatively rough surface morphologies having
either pinholes (THF) or small cross-linked particles (dioxane),
respectively. Small pinholes are also observed for films fabri-
cated from EtOAc, but possibly due to the small depths (<1
nm), they do not affect the dielectric properties of the films.
Table 3 summarizes optimized conditions (solvent and
polymer/cross-linker ratio) for thin (<20 nm) CPB dielectric
films fabricated with the various cross-linkers. Here, and for
the following studies (vide infra), dielectric film quality was
rated from excellent to poor based on the film morphological
(roughness) characteristics as well as the dielectric response
(leakage current density and breakdown characteristics). Except
for films fabricated with C6OMe, all of the new cross-linkers
afford robust CPB dielectric films. After curing at ∼110 °C,
these films are completely insoluble in the mother solvent.
Because of the very low cross-linking reactivity of C6OMe
among all of the cross-linkers investigated, robust CPB films
using C6OMe are probably formed only when cross-linking is
carried out in the presence of a Sn or Ti catalyst.24 However,
in this study, cross-linking catalysts were not employed because
of their deleterious effect on leakage current.
b. Spin-Coated Thick (50-300 nm) CPB Films. To assess
how the different cross-linker reactivities affect the CPB
dielectric property transition from thin to thicker films, similar
film optimizations based on the aforementioned cross-linking
conditions for C6Cl, EGOAc, and C6NMe2, were performed for
relatively thick CPB films (50-300 nm) on Si, Al/PEN, and
ITO/Mylar substrates. C6OMe was not studied because of the
poor performance of the corresponding thin CPB films whereas,
among the silyl acetates, C6OAc-based films exhibit character-
istics similar to those of EGOAc (vide infra). Table 4 collects
the optimized cross-linking conditions and electrical data for
the thick CPB films on Si substrates. Except for C6NMe2, thick
CPB films require different solvents to afford optimally smooth
morphologies compared to the corresponding thin films. Fur-
thermore, spin-coated film morphologies depend on the polymer/
cross-linker ratio, reagent solubility, solvent volatility/boiling
(25) (a) Lee, T.-W.; Shin, J. H.; Kang, I.-N.; Lee, S. Y. AdV. Mater. 2007,
19, 2702–2706. (b) Halik, M.; Klauk, H.; Zschieschang, U.; Kriem,
T.; Schmid, G.; Radlik, W.; Wussow, K. Appl. Phys. Lett. 2002, 81,
289–291. (c) Ullmann, A.; Ficker, J.; Fix, W.; Rost, H.; Clemens,
W.; McCulloch, I.; Giles, M. Mater. Res. Soc. Proc. 2001, 665,
C7.5.1-6.
(24) (a) Palmlof, M.; Hjertberg, T. J. Appl. Polym. Sci. 1999, 72, 521–
528. (b) Hjertberg, T.; Palmlof, M.; Sultan, B.-A. J. Appl. Polym. Sci.
1991, 42, 1185–1192.
9
6872 J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008

Printable Cross-Linked Polymer Blend Dielectrics
A R T I C L E S
Table 3. Spin-Coated CPB Thin Film (<20 nm) Deposition Conditionsa
a Polymer:crosslinker ratio (mg:mg), solvent, film quality, film thickness (D, nm), RMS roughness (F, nm), areal capacitance value (Ci, nF cm-2), and
leakage current density (J, A/cm2) at a measurement electric field of 2.0 MV/cm. Films were fabricated on n+-Si substrates. b Film morphological
quality rated as E (excellent), G (good), or P (poor). See text for explanation of terminology.
10-7 A/cm2) at a measurement electric field of 2.0 MV/cm and
have lower breakdown fields, E ) +4.0 to -4.0 MV/cm (Figure
7). EGOAc-based CPB films also afford smoother surfaces with
an rms roughness of 0.2-0.3 nm versus C6Cl (0.3-0.4 nm) or
C6NMe2 (2-3 nm)-based films. Note that even the CPB films
fabricated with C6Cl are relatively smooth, but only over small
areas since the overall film quality is poor compared to those
fabricated from EGOAc (Supporting Information, Figure S1).
Depending on the film thickness, spin-coated thick CPB films
exhibit areal capacitance values of 18-82 nFcm-2 (at 10 kHz;
Figure 6C, Table 4). Similar to the case of the spin-coated thin
films, the formal dielectric constants calculated from capacitance
values for spin-coated thick CPB films using eq 2, are reasonable
(4.7-6.2) considering the literature values.12,20,25
To investigate the compatibility of CPB films with flexible
substrates, thick CPB films were also fabricated on flexible Al/
PEN and ITO/Mylar substrates using EGOAc as the cross-linker.
Table 5 collects the optimized conditions and electrical proper-
ties of spin-coated thick CPB dielectric films on these substrates.
Compared to the films fabricated on n+-Si substrates, the CPB
dielectric films fabricated on Al/PEN or ITO/Mylar exhibit
Figure 6. Electrical properties of spin-coated thin CPB films fabricated
slightly rougher surfaces (F ) 0.5-0.7 nm) and somewhat
greater leakage current densities (∼ 1 × 10-6 A/cm2) at a
measurement electric field of 2.0 MV/cm (Figures S2 and S3).
Note that the n+-Si substrates are far smoother than Al/PEN
and ITO/Mylar, with the latter two substrates affording greater
CPB roughness.
with various cross-linkers (polymer-cross-linker ratios are as given in Table
3; C6NMe2 ) blue, C6Cl ) black, C6OAc ) red, EGOAc ) green, C6OMe
) purple). (A) Leakage current density vs voltage plot. (B) Leakage current
density vs electric field plot. Circled data indicate the region of breakdown.
(C) Capacitance-frequency plots (1-103 kHz; film deposition conditions
are given in Tables 3 and 4) for ultrathin (top) and thick (bottom) spin-
coated CPB films.
c. Printed CPB Films. Dielectric films were printed on Al/
PEN or ITO/Mylar substrates using EGOAc as the cross-linker,
which afforded the best dielectric performance among the cross-
linking reagents investigated. Printed films using C6OAc afford
comparable dielectric properties to those based on EGOAc
(Table S1, Figure S4). Similar to the optimization process carried
out for the spin-coated CPB films, experiments were performed
to identify the optimum printing solvent for a given polymer:
cross-linker precursor composition and the optimum polymer:
cross-linker ratio for a given solvent (Table 6). For a specific
PVP/EGOAc concentration ratio (100 mg:50 mg in 0.4 mL
point, and film spinning rate. Therefore, different solvents afford
films having different qualities depending on the combination
of spin-coating conditions and polymer-cross-linker reactivity.26
Similar to the thin CPB films, films fabricated using moderately
reactive EGOAc afford the smoothest film morphologies and
superior dielectric response. Thick CPB films fabricated with
C6Cl and C6NMe2afford ∼10× greater leakage current densities
(∼1 × 10-6 A/cm2) than those fabricated with EGOAc (∼1 ×
(26) Norrman, K.; Ghanbari-Siahkali, A.; Larsen, N. B. Annu. Rep. Prog.
Chem., Sect. C 2005, 101, 174–201.
9
J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008 6873

A R T I C L E S
Kim et al.
Table 4. Spin-Coated Thick CPB Film Deposition Conditions and Film Propertiesa
a Polymer:crosslinker ratio (mg:mg), solvent, film quality, film thickness (D, nm), RMS roughness (F, nm), areal capacitance value (Ci, nF cm-2), and
leakage current density at a measurement electric field of 2.0 MV/cm (J, A/cm2). Films were fabricated on an n+-Si substrate. b Film morphological
quality rated as E (excellent), G (good), or P (poor). See text for explanation of terminology.
exhibit thicknesses of 450-550 nm and areal capacitance values
of 5.5-7.5 nF cm-2. Note that the gravure-printed films afford
somewhat “wavy” surface features typical of the particular
engraved printing cylinder used in this study. Furthermore, the
gravure-printed CPB films afford very good electrical properties
with low leakage current densities of ∼1 × 10-7 A/cm2 at 2.0
MV/cm (Figure 8).
Figure S4 shows AFM images of CPB dielectric films printed
from optimized precursor compositions. These images demon-
Figure 7. Electrical properties of spin-coated thick CPB films. (PVP-C6Cl
(40:40) ) black, PVP-EGOAc (40:40) ) red, PVP-C6NMe2 (40:40) ) blue;
solvents are given in Table 4): (A) Leakage current density vs electric field
plot; (B) extended leakage current density vs electric field plot. Circled
data indicate the region of breakdown.
strate that the optimized printed dielectric films exhibit very
smooth surfaces (rms roughness ) 0.3-0.4 nm). Similarly,
Figure S5 shows optical images of gravure-printed dielectric
films fabricated using C6Cl and EGOAc cross-linkers, demon-
strating that cross-linker design is essential to achieving
acceptable printed dielectrics. Furthermore, the data of Table 6
clearly demonstrate that identifying the optimal composition of
the dielectric precursor formulations (type of polymer, solvent,
as well as polymer/cross-linker concentration ratio) is critical
to achieving optimally pinhole- and defect-free dielectric films.
Table 5. Spin-Coated Thick CPB Film Properties for Films
Fabricated from PVP-EGOAc in Dioxane (2 mL)a
entry
gate
PVP:EGOAc (mg:mg)
D
Cib
1
2
3
4
5
6
7
8
n+-Si
80:80
60:60
40:40
30:30
20:20
80:80
60:60
40:40
30:30
20:20
80:80
60:60
40:40
30:30
20:20
305
205
130
80
18
25
40
55
82
20
27
44
58
85
22
27
49
58
87
n+-Si
n+-Si
n+-Si
OFET Fabrication and Performance. Top-contact FETs were
fabricated using spin-coated and gravure-printed CPB films as
the gate dielectrics on n+-Si, Al/PEN, or ITO/Mylar gate
substrates. Both p-type (P5) and n-type (DFHCO-4T) organic
semiconductors were employed (Figure 2). FET devices were
fabricated by spin-coating or printing PVP-cross-linker mixture
solutions, curing in a vacuum oven at 110 °C, followed by
semiconductor and Au source-drain contact vapor deposition.
All OFET devices were characterized under vacuum. Transfer
characteristics of the devices were measured in the saturation
regime (VDS g VG - VT), and representative transfer plots are
shown in Figures 9–11. Tables 7 and 8 collect the OFET
performance parameters, carrier mobilities in the saturation
regime (µsat), and current on/off ratios (Ion/Ioff). µsat is calculated
from the slope of the linear part of the IDS1/2-VG plot by fitting
the data to µsat ) (2IDSL)/[WCi(VG - VT)2].11
n+-Si
50
Al/PEN
Al/PEN
Al/PEN
Al/PEN
Al/PEN
ITO/Mylar
ITO/Mylar
ITO/Mylar
ITO/Mylar
ITO/Mylar
300-310
200-210
125-135
80
9
10
11
12
13
14
15
50
300-310
200-210
125-135
80
50
a Gate dielectric, polymer:crosslinker ratio (mg:mg), film thickness
(D, nm), and areal capacitance value (Ci, nF/cm2). b Measured in a MIS/
MIM structure.
solvent), films fabricated with EtOAc as the solvent afford by
far the most impressive dielectric characteristics. Similarly, the
optimum polymer/cross-linker concentration ratios were deter-
mined to be PVP:cross-linker at ∼100 mg:50 mg in 0.4 mL
solvent, using EtOAc as the solvent and using the printing
parameters provided in the Experimental Section (see Supporting
Information). Gravure-printed CPB films fabricated under
optimum printing conditions (solvent, polymer/cross-linker ratio)
a. OFETs Fabricated with Spin-Coated CPB Films. Figures
9 and 10 show representative transfer and output plots for P5
and DFHCO-4T transistors fabricated with spin-coated CPB
films as gate dielectrics. As shown in Figure 9, devices
fabricated with thin CPB dielectric films function at low
operating voltages ((3.0 V) as a consequence of the large areal
9
6874 J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008

Printable Cross-Linked Polymer Blend Dielectrics
A R T I C L E S
Table 6. Gravure-Printed CPB Dielectric Film Fabrication Conditions and Film Propertiesa
a Polymer:crosslinker ratio (mg:mg), solvent, film quality, film thickness (D, nm), RMS roughness (F, nm), areal capacitance value (Ci, nF cm-2), and
leakage current density at a measurement electric field of 2.0 MV/cm (J, A/cm2). Films were fabricated on Al/PEN or ITO/Mylar substrate. b Film
morphological quality rated as E (excellent), G (good), or P (poor). See text for explanation of terminology.
Figure 8. Electrical properties of gravure-printed CPB films: (A) leakage
current density vs electric field plot; (B) capacitance-frequency plots
(10-103 kHz) for gravure-printed CPB films. (PVP:EGOAc (100:50),
polymer (mg):cross-linker (mg) in 0.4 mL EtOAc). Arrow indicates an areal
capacitance range of 5.5-7.5 nF cm-2
.
capacitance values (300-350 nF/cm2). Device operating volt-
ages increase as the gate dielectric film thicknesses increase (as
the capacitance values decrease) from 3.0 V (PVP-EGOAc 4:6,
14 nm thickness) to 60 V (PVP-EGOAc 80:80, 305 nm
thickness). These OFETs exhibit mobilities of 0.12-0.49 and
0.02-0.52 cm2/Vs for P5 and DFHCO-4T, respectivelyscom-
parable to those based on 300 nm thick SiO2 dielectrics (Table
7). Note that because of the optimized gate dielectric sur-
face morphologies, P5 devices afford higher mobilities than
those achieved with first-generation CPBs (0.10 cm2/Vs).12
These results demonstrate the importance of cross-linker and
solvent optimization. For devices fabricated with all of the
present dielectric films, Ion/Ioff ) 104-106. The high contact
resistance observed for the output characteristics of DFHCO-
4T-based devices (Figures 9D and 10F) can be attributed to
the high injection barrier between the DFHCO-4T LUMO (3.96
eV)21 and the Au work function (5.1 eV). OFETs fabricated on
flexible Al/PEN substrates afford somewhat lower performance
compared to those on Si substrates with µsat ) 0.26-0.34 cm2/
Vs and Ion/Ioff ) ∼105 (Table 7, Figure S6).
Figure 9. Performance of OFET devices fabricated with spin-coated thin
CPB films using various cross-linkers (polymer/cross-linker ratios are as
given in Table 3: C6NMe2 ) blue, C6Cl ) black, C6OAc ) red, EGOAc
) green). OFET transfer characteristics for (A) P5, (B) DFHCO-4T; OFET
output characteristics for (C) P5 on PVP-EGOAc (4:6) and (D) DFHCO-
4T on PVP-C6OAc (4:6). The channel width and length are 500 and 100
µm (for P5) or 1000 and 100 µm (for DFHCO-4T), respectively.
found to exhibit good device performance metrics, affording
carrier mobilities of 0.3-0.9 cm2/Vs. Figure 11 shows repre-
sentative transfer and output plots of pentacene film OFET
devices fabricated on one of the gravure-printed PVP-EGOAc
(100:50) formulations. Note also that the printed OFET devices
exhibit excellent mechanical flexibility (Figure 12).
b. OFETs Based on Printed CPB Films. Similar to the spin-
coated films, gravure-printed CPB films on Al/PEN or ITO/
Mylar substrates were also employed as dielectrics for OFET
fabrication. Table 8 summarizes OFET performance data for
the two different semiconductors (P5 and DFHCO-4T) on
various gravure-printed CPB films. In all cases, the devices are
FET-active and the computed carrier mobilities are comparable
to those on 300 nm-thick SiO2 substrates (Table 7). P5 and
DFHCO-4T transistors fabricated on the printed dielectric are
Discussion
CPB Morphology versus FET Response Characteristics. All
cross-linking reactions employed in this study yield robust,
insoluble CPB siloxane networks (Scheme 1), exhibiting identi-
cal chemical composition. The only major differences between
the various spin-coated thin CPB films are the details of film
morphologies (roughness), originating from the differing cross-
9
J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008 6875

A R T I C L E S
Kim et al.
Table 7. Carrier Mobility (µsat, cm2/Vs) and Current On/Off Ratio
(Ion/Ioff) Data for OFET Devices Fabricated Using P5 and
DFHCO-4T as the p- and n-type Organic Semiconductor on
Various CPBsa
with C6OAc and EGOAc afford large, dendritic grains (>2 µm),
whereas far smaller grains (<0.4 µm) are observed on C6Cl-
and C6NMe2-based dielectric films (Figure 13). Note that P5
films deposited on spin-coated thick EGOAc-based CPB films
(D ) 50-305 nm) afford large, dendritic grains, independent
of the dielectric film thickness. Because of the smooth film
morphology, these P5 FETs exhibit similar carrier mobilities
(Table 7, Figure 13).
The high carrier mobility and current on/off ratio achieved
with P5 and DFHCO-4T FETs based on spin-coated and
gravure-printed CPB films also imply very good semiconductor
film crystallinity. Figures 14 and S7 show XRD θ-2θ scans of
50 nm-thick P5 and DFHCO-4T films deposited on spin-coated
and gravure-printed CPB films. Indeed, all of the XRD spectra
exhibit sharp, multiple Bragg reflections demonstrating a large
degree of texture, with those fabricated on smoother dielectric
surfaces exhibiting far larger intensities. It is well-known that
rough surfaces disrupt the growth of highly textured pentacene
grains.27a,29 A similar effect of the dielectric morphology on
semiconductor film growth is observed for the first time here
for DFHCO-4T.
P5
DFHCO-4T
b
µsat
b
µsat
substrate
cross-linker
ratio
Ion/Ioff
Ion/Ioff
n+-Si
C6NMe2
C6Cl
C6OAc
EGOAc
4:6
6:6
4:6
4:6
0.12
0.18
0.35
0.37
105
105
104
104
105
106
106
107
105
105
105
104
105
0.02
0.11
0.12
0.16
0.44
0.43
0.47
0.52
105
104
106
105
106
106
105
105
20:20 0.40
40:40 0.34
60:60 0.43
80:80 0.49
20:20 0.30
40:40 0.34
60:60 0.26
0.10
Al/PEN EGOAc
n+-Si
n+-Si
first generation CPBc
d
300 nm SiO2
0.39
0.37
107
a Gate dielectric films were spin-coated on n+-Si or Al/PEN
substrates. b Calculated for the charge carrier concentration nQ ) CiVG/e
) 5-6 × 1012 cm-2
.
c From previous study (ref 11). d Data are for 300
nm-thick SiO2/Si devices with pentacene films grown under the same
conditions as for other CPB films.
Humidity-Dependent Electrical Properties. Previous studies
showed that the performance of OFETs based on PVP or cross-
linked PVP gate dielectric films are significantly affected by
surface polarization at the semiconductor/dielectric interface
induced by absorbed water molecules in the dielectric.30 Thus,
OFETs fabricated with the present CPB materials were char-
acterized both in vacuo and in humidity-controlled environ-
ments. To provide comprehensive characterization of the CPB
materials, humidity-dependent dielectric properties and the
corresponding P5 FET performance were investigated for the
different CPB compositions. For this study, only CPB films
affording excellent surface characteristics such as spin-coated
PVP-EGOAc (4:6, D ) 14 nm), PVP-EGOAc (40:40, D ) 130
nm), and PVP-EGOAc (80:80, D ) 305 nm) and printed PVP-
EGOAc (100:50, D ) 450 - 550 nm; as given in Tables 3, 4,
and 6) were employed. Figure 15 shows measured leakage
current densities plotted versus electric field under various
humidity levels (from vacuum to relative humidity (RH) )
36%). In all cases, the leakage current density of these MIS/
MIM capacitors increases when the RH is increased. In the case
of the spin-coated thin CPB films (PVP-EGOAc 4:6) with the
lowest PVP content, the leakage current densities at high RH
(>26%) increase about an order of magnitude at most, compared
with those measured at low RH (<8%). As the PVP content
increases, going from thin to thick spin-coated films to printed
CPB films, the leakage current densities become more sensitive
to moisture. Thus, for spin-coated thick CPB films, J increases
slightly at RH ) 8% compared to vacuum, and to >10× at RH
> 26%. Furthermore, printed CPB films having the greatest PVP
content afford J ≈ 10× greater at RH ) 8%, and J > 100×
greater at RH > 26%.
Table 8. Carrier Mobility (µsat, cm2/Vs) and Current On/Off Ratio
(Ion/Ioff) for OFET Devices Fabricated with Various Gravure-Printed
CPB Films on Al/PEN and ITO/Mylar substrates
PVP (mg):EGOAc
(mg) (0.4 mL EtOAc)
substrate
semiconductor
µsat
Ion/Ioff
Al/PEN
ITO/Mylar
Al/PEN
100:50
100:50
100:50
100:50
P5
P5
0.29-0.76
0.44-0.85
0.58-0.86
0.38-0.60
106
104
105
106
DFHCO-4T
DFHCO-4T
ITO/Mylar
linker reactivities. Note that as a consequence of the uniformly
low leakage current densities (<1 × 10-6 A/cm2 at 2.0 MV/
cm) for all spin-coated CPB films, the differing dielectric
properties of the various films have little effect on OFET
responses at relatively low device operating voltages. However,
the OFET performance trends measured for devices fabricated
with the spin-coated CPB films are rather sensitive to the
dielectric film morphologies. For spin-coated thin CPB films,
the P5 FET devices fabricated with the EGOAc-based dielec-
trics, having the smoothest film morphology, exhibit the largest
carrier mobility, ∼0.40 cm2/Vs, among all cross-linkers exam-
ined, followed by C6OAc (µsat ) 0.35 cm2/Vs), C6Cl (µsat
)
0.18 cm2/Vs), and C6NMe2 (µsat ) 0.12 cm2/Vs). Interestingly,
a similar dielectric morphology-mobility trend is observed for
the n-type semiconductor DFHCO-4T. Since OFET charge
transport is confined to the nanoscopic region at the semicon-
ductor/dielectric interface, the smoothness of the dielectric film
surface is clearly a prerequisite here to optimizing OFET
performance. In the case of pentacene FETs, several groups23,27
have reported that rough gate dielectric surfaces induce the
growth of smaller pentacene grains rather than with very smooth
substrates, and that roughness leads to poor FET response. Grain
boundaries between semiconductor crystallites are thought to
function as interfacial charge trapping sites which disrupt carrier
drift.28 Pentacene films grown on smoother CPB films fabricated
P5 OFETs fabricated with the present CPB films exhibit
similar behavior when evaluated under conditions of differing
humidity. Spin-coated thin CPB films (PVP-EGOAc 4:6) are
least affected by the various humidity levels. For these films,
field-effect mobilities afford little variation between vacuum and
RH ) 26%, and a slight increase at RH ) 36% (not shown).
(27) (a) Kim, C.; Facchetti, A.; Marks, T. J. AdV. Mater. 2007, 19, 2561–
2566. (b) Steudel, S.; de Vusser, S.; de Jonge, S.; Janssen, D.; Verlaak,
S.; Senoe, P.; Heremans, J. Appl. Phys. Lett. 2004, 85, 4400–4402.
(28) Bolognesi, A.; Berliocchi, M.; Manenti, M.; Carlo, A. D.; Lugli, P.;
Lmimouni, K.; Dufour, C. IEEE Trans. Electron DeVices 2004, 51,
1997–2003.
(29) Kim, C.; Facchetti, A.; Marks, T. J. Science 2007, 318, 76–80.
(30) (a) Jung, T.; Dodabalapur, A.; Wenz, R.; Mohapatra, S. Appl. Phys.
¨
Lett. 2005, 87, 182109/1-3. (b) Ba¨cklund, T. G.; Osterbacka, R.;
Stubb, H.; Bobacka, J.; Ivaska, A. J. Appl. Phys. 2005, 98, 074504/
1-6.
9
6876 J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008

Printable Cross-Linked Polymer Blend Dielectrics
A R T I C L E S
Figure 10. Performance of OFET devices fabricated with spin-coated thick CPB films (deposition conditions are as given in Table 4). OFET transfer
characteristics for P5 on (A) PVP-EGOAc (20:20), (B) PVP-EGOAc (40:40) and for DFHCO-4T on (C) PVP-EGOAc (60:60), and (D) PVP-EGOAc
(80:80). OFET output characteristics for (E) P5 on PVP-EGOAc (40:40), and (F) DFHCO-4T on PVP-EGOAc (80:80). The channel width and length are
500 and 100 µm (for P5) or 1000 and 100 µm (for DFHCO-4T), respectively.
Figure 11. Transfer and output characteristics for P5 OFET devices
fabricated on gravure-printed CPB (PVP-EGOAc ) 100 mg:50 mg in 0.4
Figure 13. AFM images of 50 nm thick pentacene films grown on various
mL EtOAc) dielectric films using ITO/Mylar as the substrate. The channel
spin-coated thin and thick CPB films (PVP (mg)/cross-linker (mg) in 2 mL
length and width are 100 and 500 µm, respectively.
of solvent as given in Tables 3 and 4). The scale bars indicate 1 µm.
significant for OFET devices based on printed CPB gate
dielectric films. Thus, P5 OFET mobilities increase from 0.36
cm2/Vs under vacuum, to 3.3 at RH ) 8%, to 5.3 at RH )
26% and 8.5 at RH ) 36%. The increase in source-drain current
is attributed to the accumulation/stabilization of extra charge
carriers at the semiconductor/dielectric interface induced by the
surface polarization from the absorbed moisture in the dielectric
film.30 At high humidities, due to increased gate leakage current,
decreased source-drain current is observed at high gate voltages.
The previously reported exceptional mobilities (3-5 cm2/Vs)
of P5 OFET devices using PVP or PVP-like polymers as
dielectrics could also be attributed to such effects.1c,31 Because
of an increased source-drain current in the off-state, Ion/Ioff
decreases at high RH for all OFET devices examined (Figure
16).
Figure 12. Photographs of the indicated printed OFETs demonstrating
mechanical flexibility.
Conclusions
Figure 16 shows P5 OFET transfer characteristics for devices
fabricated with spin-coated (PVP-EGOAc 40:40) and printed
(PVP-EGOAc 100:50) CPB films. For the spin-coated CPB
films (Figure 16A,B), the field-effect mobility of P5 OFETs
increases slightly from 0.35 to 0.50 cm2/Vs when the moisture
level increases from RH ) 0% (vacuum) to RH ) 26%. An
even greater µ of 0.90 cm2/Vs is observed at RH ) 36%. The
mobility increase at higher moisture levels is even more
We have reported newly designed R,ω-disilane cross-linkers
with tuned reactivity, optimized by an in situ reaction kinetic
study. With these new cross-linkers, we have demonstrated that
optimization of cross-linking processes with PVP, employing
(31) Lee, S.; Koo, B.; Shin, J.; Lee, E.; Park, H.; Kim, H. Appl. Phys.
Lett. 2006, 88, 162109/1-3.
9
J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008 6877

A R T I C L E S
Kim et al.
Figure 14. The θ-2θ XRD spectra of 50 nm-thick P5 (left) and DFHCO-4T (right) films grown on spin-coated CPB films (PVP/cross-linker ratios are as
indicated).
Figure 15. Leakage current densities versus electric field of various CPB
Figure 16. Response of P5 OFET devices fabricated with CPB films under
different relative humidity levels (black, vacuum; blue, RH 8%; red, RH
26%; green, RH 36%). Transfer characteristics on spin-coated thick (PVP-
films measured under different relative humidity (RH) levels (black, vacuum;
blue, RH 8%; red, RH 26%; green, RH 36%): (A) spin-coated thin films
(PVP-EGOAc, 4:6); (B) spin-coated thick films (PVP-EGOAc 40:40); (inset)
expanded leakage current density plot. Circled data indicate the region of
breakdown; (C) spin-coated thick films (PVP-EGOAc 80:80). For spin-
coated films, PVP (mg)/EGOAc (mg) in 2 mL of dioxane. (D) Printed films
(PVP-EGOAc 100:50 on Al/PEN film). For printed films, PVP (mg)/EGOAc
(mg) in 0.4 mL of EtOAc.
1/2
EGOAc 40:40) films: (A) IDS versus VG; (B) IDS versus VG. Transfer
characteristics on printed (PVP-EGOAc 100:50 on Al/PEN substrate) films:
(C) IDS versus VG; (D) IDS1/2 versus VG. Channel width and length are 1000
and 100 µm (A and B) or 500 and 100 µm (C and D), respectively. VDS is
-30 and -80 V for spin-coated and printed films, respectively.
moderately reactive cross-linkers, appropriate solvent, and
appropriate polymer/cross-linker concentration ratios affords
robust, adherent, pinhole-free, and smooth gate dielectric
materials with tunable thickness and capacitance. These films
are readily deposited from solution by spin-coating or gravure
printing, adhere strongly to a variety of conducting substrates,
and are compatible with p- and n-type organic semiconductors.
These results demonstrate that robust, low-leakage polymer
dielectric films are accessible and that the FET devices utilizing
these polymer dielectrics in solution-processed fabrication
methodologies offer attractive opportunities for printing ap-
plications. While the hygroscopic nature of PVP to some degree
affects electrical properties of the dielectric films and the
resultant OFET device performance (depending on the exact
film composition), such effects are not in general severely
detrimental to OFET performance and should be suppressable
by simple encapsulation procedures.32 Moisture absorbed in the
dielectric film appears to induce additional charge carriers,
resulting in increased gate leakage and source-drain current as
well as in apparent mobility.
Acknowledgment. This contribution is dedicated to the memory
of Mr. Hyuk-Jin Choi. This research was supported by ONR (Grant
N00014-02-0909) and Polyera Corp. We thank the NSF-MRSEC
program through the Northwestern Materials Research Center
(Grant DMR-0520513) for providing characterization facilities.
Supporting Information Available: Experimental details,
optical microscope image, AFM image, XRD, leakage current
density, and OTFT device performance data for chosen dielectric
materials. This material is available free of charge via the
Internet at http://pubs.acs.org.
(32) (a) Feili, D.; Schuettler, M.; Doerge, T.; Kammer, S.; Stieglitz, T.
Sensors and Actuators A 2005, 120, 101–109. (b) Kawashima, N.;
Nomomto, K.; Wada, M.; Kasahara, J. Mater. Res. Soc. Proc. 2005,
871, I1.10.1-6.
JA801047G
9
6878 J. AM. CHEM. SOC. VOL. 130, NO. 21, 2008