High electron mobility in nickel bis(dithiolene) complexes

Article abstract of DOI:10.1039/b701036b

The charge-carrier mobilities for three Ni bis(dithiolene) complexes have been determined using the steady-state space-charge limited current technique. A high mobility of 2.8 cm² V^-1 s^-1 was observed for one compound, which exhibits a π-stacked columnar structure, in an annealed unsymmetrical melt-processed device. Energy-level considerations and field-effect transistor measurements suggest that this value represents an electron mobility. However, saturation mobilities measured for this compound in spin-coated field-effect transistors were found to be over two orders of magnitude lower than the space-charge limited current values. X-Ray diffraction shows a difference in morphology between thick melt-processed and thin spin-coated films and, therefore, a significant change in intermolecular packing between the device types may explain the discrepancy in mobilities obtained using the two techniques. The Royal Society of Chemistry.

Full text of DOI:10.1039/b701036b

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
High electron mobility in nickel bis(dithiolene) complexes
Jian-Yang Cho,abd Benoit Domercq,ac Simon C. Jones,ab Junsheng Yu,ac Xiaohong Zhang,ac Zesheng An,ab
Maximilienne Bishop,d Stephen Barlow,*abd Seth R. Marder*abd and Bernard Kippelen*ac
Received 23rd January 2007, Accepted 22nd March 2007
First published as an Advance Article on the web 18th April 2007
DOI: 10.1039/b701036b
The charge-carrier mobilities for three Ni bis(dithiolene) complexes have been determined using
the steady-state space-charge limited current technique. A high mobility of 2.8 cm2 V21 s21 was
observed for one compound, which exhibits a p-stacked columnar structure, in an annealed
unsymmetrical melt-processed device. Energy-level considerations and field-effect transistor
measurements suggest that this value represents an electron mobility. However, saturation
mobilities measured for this compound in spin-coated field-effect transistors were found to be
over two orders of magnitude lower than the space-charge limited current values. X-Ray
diffraction shows a difference in morphology between thick melt-processed and thin spin-coated
films and, therefore, a significant change in intermolecular packing between the device types may
explain the discrepancy in mobilities obtained using the two techniques.
self-exchange rates in solution for many metal–organic redox
Introduction
systems, suggesting high ET rates might also be obtainable
in films.4–7 Nevertheless, TM complexes have been rather
Solution- or vapor-processible molecular or polymeric mate-
rials with high charge-carrier mobilities are critical to the
underexplored. Several studies have focused on the effect of
development of efficient organic light-emitting diodes
incorporating metals into the main chains of conjugated
polymers,8–10 while Thompson and co-workers have used
(OLEDs), photovoltaic cells, and field-effect transistors
(FETs).1–3 The hopping-type conduction mechanism in
semiconductors (i.e. for most organic electronic materials)
can be regarded as a series of electron-transfer (ET) reactions
between the neutral species and its radical cation (in the case of
hole-transport materials) or radical anion (electron-transport).
According to Marcus theory, the barrier to such ‘‘self-
exchange’’ ET reactions, DG{, depends principally on the
reorganization energy, l, while the pre-exponential factor
influencing the ET rate increases with the intermolecular
organocobalt species as hole-transport materials in OLEDs;11
however, studies of carrier mobilities of neutral molecular
d-metal complexes remain scarce. Mobilities have been mea-
sured in d- and f-block phthalocyanine complexes, although in
most cases the redox processes involved in charge transport are
likely to involve orbitals substantially localized on the ligands
rather than the metal centers.12 Re(I) and Ru(II) 2,29-bipyridyl
complexes and a Re(I) diimine complex have been shown to
exhibit bipolar charge transport character with both electron
and hole mobilities in the order of 1024 to 1023 cm2 V21 s21
measured using the time-of-flight method (TOF).13,14
electronic coupling, HAB
.
The vast majority of molecular and polymeric species used
as transport materials to date have been based upon the lighter
p-block elements. However, transition-metal (TM) compounds
have several attractive features for charge transport applica-
tions; often frontier orbitals are more-or-less non-bonding
(suggesting low reorganization energies), yet with some
delocalization onto the ligands (favoring intermolecular
overlap). Moreover, redox processes are commonly fully
reversible (indicating that the charge-bearing species are
reasonably kinetically stable), with potentials that can be
tuned over a wide range through variation of metal and/or
ligand. Importantly, there exists empirical evidence for high
Very recently, FET hole mobilities as high as 3.8 6
1022 cm2 V21 s21 have been reported for Ni bis(o-diimino-
benzo-semiquinonate) complexes such as I15 (Fig. 1) and a
methyl-substituted analogue.16 Ni bis(dithiolene) complexes
have also recently been used to fabricate FETs with electron
mobilities ranging from 3.0 6 1026 to 2.0 6 1025 cm2 V21 s21
for II16 and from ca. 1024 to ca. 1023 cm2 V21 s21 for III17
(Fig. 1). Compound III was also found to exhibit comparable
FET hole mobility and the transistors have been used to
fabricate complementary circuits.17
aCenter for Organic Photonics and Electronics (COPE), Georgia
Institute of Technology, Atlanta, Georgia, 30332, USA
bSchool of Chemistry and Biochemistry, Georgia Institute of
Technology, Atlanta, Georgia, 30332, USA.
E-mail: stephen.barlow@chemistry.gatech.edu;
seth.marder@chemistry.gatech.edu; Fax: (+1) 404 894 5909
cSchool of Electrical and Computer Engineering, Georgia Institute of
Technology, Atlanta, Georgia, 30332, USA.
E-mail: kippelen@ece.gatech.edu; Fax: (+1) 404 385 5170
dDepartment of Chemistry, University of Arizona, Tucson, Arizona,
85721, USA
Fig. 1 Square-planar Ni complexes previously studied as transport
materials.
2642 | J. Mater. Chem., 2007, 17, 2642–2647
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Attractive features of Ni bis(dithiolene) complexes for
applications as molecular transport materials18–22 include:
extensive delocalization of their frontier molecular orbitals
over the entire complex (suggesting the possibility of
significant intermolecular overlap in films, and low reorgani-
zation energies upon reduction and/or oxidation); thermal and
photochemical stability; widely tunable and reversible redox
chemistry (often forming chemically isolable anions); small
structural changes observed on reduction (suggesting low
reorganization energies for this process); and an approximately
planar shape (offering the possibility of self-assembly into
columnar mesophases). In addition to the recent work
mentioned above, which shows that neutral Ni bis(dithiolene)
complexes can act as ET materials in FETs, the potential for
substantial intermolecular overlap and, consequently, for
interesting electronic behavior has been shown by previous
work on salts of reduced metal bis(dithiolene) anions; for
example, many species of [X]x[M(dmit)2] (X = organic cation,
dmit = 1,3-dithiol-2-thione-4,5-dithiolato) contain stacked
bis(dithiolene) units and are metallic, [Me4N]0.5[Ni(dmit)2] is
a superconductor under pressure,23 and an electron mobility
of ca. 0.2 cm2 V21 s21 has been measured in an FET
incorporating [N-octadecylpyridinium][Ni(dmit)2].24 Here we
report on investigations of charge transport in neutral Ni
bis(dithiolene) complexes 1, 2, and 3 (Fig. 2).
X-ray diffraction data at 150 uC (Fig. 3a) reveal a solitary,
intense peak at low angle, indicating this mesophase to be
˚
lamellar, with 43.9 A interlayer spacing. In contrast, 3 displays
no LC behavior, melting at 219 uC. All three species exhibit
reversible one-electron molecular oxidations and reductions in
dichloromethane; the potentials are summarized in Table 1,
along with condensed-phase HOMO and LUMO energies
estimated by comparison of the potentials with those of com-
pounds for which ionization potentials and electron affinities
have been measured directly.28–30 These values may be com-
pared to 5.2 and 4.3 eV previously estimated for the HOMO
and LUMO energies of III from electrochemical data.17
Results and discussion
Complex 1 was synthesized as described by Ohta and co-
workers, who reported that it exhibits a hexagonal columnar
discotic liquid crystalline (LC) mesophase between 72–
108 uC.25–27 Complexes 2 and 3 are new compounds and were
synthesized from the corresponding diketones as described in
the Experimental. Polarized optical microscopy (POM) and
differential scanning calorimetry (DSC) measurements are con-
sistent with 2 forming an LC mesophase between 140–220 uC;
Fig. 3 (a) X-Ray diffraction data for mesophase of 2 measured at
150 uC, and POM photograph of SCLC device fabricated with 2
(inset); (b) X-ray diffraction data for films of 1 in SCLC (top, solid
line) and FET (bottom, dotted line) devices (offset for clarity), with
corresponding POM photograph of SCLC device (inset). Image sizes:
250 mm 6 160 mm.
Fig. 2 Syntheses and structures of the Ni bis(dithiolene) complexes
studied in this work.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 2642–2647 | 2643

View Article Online
Table 1 Electrochemical potentials (V, vs. FeCp2+/0, in CH2Cl2–0.1 M
[nBu4N][PF6]) and estimated solid-state orbital energiesa (eV) for nickel
bis(dithiolene) complexes
The material showing the highest mobility, 1, was further
tested in an unsymmetrical device configuration (ITO–1
(5 mm)–Ag). The Ag electrode has a lower workfunction
than ITO, and so more efficient electron injection might be
anticipated from an Ag contact. As shown in Fig. 4, the J–V
characteristics of this device are strongly asymmetric, as
anticipated from the differences in workfunction between the
two contacts. The highest current densities are measured when
Ag is the cathode, consistent with more facile injection of
electrons into the LUMO level of 1. With this increased
injection efficiency from the Ag electrode, the effective
+/0
E1/2
0/2
E1/2
2/22
E1/2
Compound
EHOMO
ELUMO
1
2
3
a
+0.46
+0.61
+0.61
20.55
20.51
20.48
21.34
21.30
21.28
5.6
5.8
5.8
4.1
4.1
4.1
Calculated as described in reference 28.
The charge-carrier mobilities of 1, 2 and 3 were studied by
the steady-state space-charge limited current (SCLC) techni-
que.31–34 For these measurements, 5 mm-thick samples were
fabricated by melting each compound between two glass slides
coated with conductive electrode materials. Melting was
achieved on a hot plate in air at temperatures of 120, 230
and 220 uC for 1, 2 and 3 respectively. Devices with injecting
contacts exhibit nonlinear current–voltage characteristics that
are composed of a linear ohmic region at low current densities
(J) and a nearly quadratic region (see inset of Fig. 4), due to
space-charge effects, at higher J. Effective mobility values can
be extracted from fitting the nearly quadratic re´gime to a
modified Mott–Gurney equation.32,34 When using indium tin
oxide (ITO) as the injecting contact, the effective carrier
mobilities of 1 and 2 were determined to be 0.85 and 2.5 6
1023 cm2 V21 s21, respectively, while the effective carrier
mobility determined for complex 3 using the same method was
mobility derived from an SCLC analysis is 2.8 cm2 V21 s21
.
We have also investigated the utility of 1 in an FET device
with an Ag source and drain electrodes. Current–voltage and
transfer characteristics of a device with a channel width of
2000 mm and length of 100 mm are shown in Fig. 5. For the
current–voltage characteristics shown in Fig. 5a, the gate-
source voltage VGS was increased from 0 to 60 V in 20 V steps.
When operated as n-channel transistors, these devices exhibit
characteristic saturation behavior, whereas under negative gate
bias the transistors did not operate in p-channel mode,
confirming that injection of electrons from the electrodes into
1 is the main origin of the current densities measured in the
SCLC measurements described above (FETs based on III
function as both n and p-channel transistors, but (i) the
estimated HOMO of III is higher and (ii) Au, rather than Ag,
contacts were used). Using eqn (1), see Experimental, an
effective mobility value of 1.3 6 1023 cm2 V21 s21 was
determined to be much smaller, ca. 4.5 6 1025 cm2 V21 s21
.
Although the SCLC technique does not allow us to distinguish
whether these values represent electron or hole mobilities,
energetic considerations based on the estimated orbital
energies in Table 1, and additional experiments described
below, suggest they are more likely to be electron mobilities.
The higher mobilities determined for 1 and 2 are potentially
due to the greater degree of intermolecular order present
in these two materials; POM images of the corresponding
devices (see Fig. 3 insets) show that room-temperature films
of 1 and 2 exhibit optical textures similar to those seen in
their LC phases, while, under the same conditions, films
of 3 show no texture, consistent with these samples being
amorphous.
calculated from the transfer characteristic shown in Fig. 5b. A
pffiffiffiffiffiffiffi
linear fit to the IDS vs. VGS data yielded a threshold voltage of
4.6 V and semi-logarithmic plots of IDS vs. VGS (IDS = drain-source
current) yielded an on–off current ratio of .105 and
a
Fig. 5 (a) Current–voltage characteristics (IDS vs. VDS) as the gate
Fig. 4 J–V characteristics of an ITO–1 (5 mm)–Ag device; positive
bias corresponds to the Ag contact as the cathode. The inset shows a
log–log plot of the data for positive bias; the solid line represents
Ohm’s law and the dashed line represents the SCLC model.
voltage (VGS) is increased from 0–60 V in 20 V steps (bottom to top);
pffiffiffiffiffiffiffi
(b) transfer characteristics (IDS and IDS vs. VGS) for a bottom-
contact FET device using 100 nm of 1 with W = 2000 mm, L = 100 mm.
2644 | J. Mater. Chem., 2007, 17, 2642–2647
This journal is ß The Royal Society of Chemistry 2007

View Article Online
sub-threshold slope of 1.5 V decade21. Mobility values were found
to decrease (in general) with shorter channel length, with typical
values 1.0 6 1024 cm2 V21 s21 in devices with a channel width of
2000 mm and length of 25 mm.
diffraction data were collected using a h-h Scintag powder
diffractometer (Cu Ka radiation) in the diffraction angle (2h)
range between 2u and 30u in 0.02u steps. Samples of 1 were
prepared on a copper plate and measured at 150 uC, while
samples of 2 were prepared on ITO-coated glass substrates by
melt-processing (SCLC devices) or Si wafers by spin-coating
(FET devices) as described below and measured at room
temperature. 2-D unit cell parameters for the SCLC sample of
1 were obtained from the first five reflections observed (after
indexing), using a least-squares refinement procedure in the
program Chekcell.35 Electrochemical measurements were
carried out under argon on dry, deoxygenated dichloro-
methane (distilled from CaH2) solutions, ca. 1024 M in
The mobility values measured in FET geometry for 1 are
comparable to, or even slightly exceed, those measured in
previously studied Ni bis(dithiolene) complexes.16,17 However,
they are several orders of magnitude lower than that measured
for 1 using SCLC. X-Ray diffraction measurements (Fig. 3b)
reveal that samples of 1 in melt-processed SCLC and spin-
coated FET devices differ in morphology and, consequently, in
the degree of long-range intermolecular order present in the
films. Samples of 1 in the SCLC devices show many Bragg
peaks; a single strong reflection is seen at low angle, followed
by a series of weaker peaks that, together, may be indexed to a
simple rectangular 2-D unit cell (unit cell parameters 34.9 6
analyte and 0.1
M
in [nBu4N][PF6], using
a
BAS
Potentiostat, a glassy carbon working electrode, a Pt auxiliary
electrode, and, as a pseudo-reference electrode, an Ag wire
anodized in 1 M KCl (aq).
˚
14.2 A), suggesting that a columnar arrangement of discotic Ni
dithiolene cores is retained in this substantially crystalline
sample. The cell a-parameter is approximately consistent with
the core–core distance determined for bulk samples of 1 in its
Synthesis and characterization
hexagonal columnar discotic LC phase (unit cell 35.6 A);26 the
˚
Nickel bis[1,2-di(39,49-didodecyloxyphenyl)ethane-1,2-dithio-
lene] (1). A mixture of 3,39,4,49-tetra-n-dodecyloxybenzil27 (S1)
(3.7 g, 3.9 mmol), phosphorus pentasulfide (3.85 g, 17 mmol)
in 1,4-dioxane (100 mL) was refluxed at 110 uC for 5 h under
an N2 atmosphere. The reaction mixture was allowed to cool
down to ca. 60 uC. The hot reaction mixture was filtered under
nitrogen through a sintered glass frit, to remove unreacted
yellow solid. NiCl2?6H2O (0.52 g, 2.19 mmol) in H2O (1.5 mL)
was added to the filtrate and the resulting mixture was refluxed
for an additional 2 h. The reaction mixture was allowed to cool
down to room temperature overnight. A grayish-blue solid
precipitated out of the reaction mixture and was collected by
filtration. The crude product was washed with acetone and
ethanol and purified by column chromatography (silica gel,
dichloromethane : hexanes = 3 : 1) to give a grayish-blue solid.
The solid was again washed with acetone and ethanol and
dried under vacuum to give a grayish-blue solid (0.89 g,
b-parameter is considerably smaller, indicating that the
columns adopt a tilted arrangement in the unit cell, with
the normal to the molecular plane inclined at ca. 66u to the
columnar stacking direction. This assignment is in agreement
with POM images of the SCLC devices, which display the fan-
shaped texture characteristic of a columnar discotic LC phase
(Fig. 3b inset). Presumably, since the sample is microcrystal-
line, columnar stacks are arranged in all possible directions
on the substrate (although not necessarily isotropically).
Accordingly some columns will be aligned with the p-stacking
between Ni dithiolene cores oriented perpendicular to the
substrate, affording favorable pathways for charge transport
between the electrodes and allowing for the high mobilities
observed in SCLC measurements. In contrast, X-ray diffrac-
tion shows that 1 adopts a lamellar structure in the FET
˚
devices studied, with a 31.9 A interlayer spacing indicating that
1
the molecular planes lie at ca. 26u to the stacking direction.
Thus, while there is likely to be some intralayer p-overlap, the
long-range columnar ordering observed for the SCLC devices
is not present in the FET devices. These changes in
morphology are attributed to the differences in processing
used to fabricate the thick and thin films used for SCLC and
FET experiments, respectively.
29.8%). H NMR (300 MHz, CDCl3) d 7.00 (dd, J = 8.2 Hz,
1.9 Hz, 4H), 6.84 (d, J = 1.9 Hz, 4H), 6.75 (d, J = 8.5 Hz, 4H),
3.96 (t, J = 6.6 Hz, 8H), 3.75 (t, J = 6.6 Hz, 8H), 1.14–1.90 (m,
160H), 0.86 (t, J = 7.1 Hz, 12H), 0.85 (t, J = 7.1 Hz, 12H).
13C{1H} NMR (75 MHz, CDCl3) d 180.38, 149.81, 148.39,
134.33, 121.85, 114.19, 112.69, 69.09, 32.01, 29.84, 29.78,
29.75, 29.54, 29.48, 29.28, 29.13, 26.11, 22.78, 14.22. IR
(cm21): 2923, 2853, 1594, 1514, 1356, 1260, 1139. UV (CHCl3)
l
max/nm (e/M21 cm21): 310 (57450), 348 (sh, 31395), 572
Experimental
(3680), 638 (sh, 3360), 962 (35210). HRMS-MALDI (m/z):
[M]+ calcd for C124H212NiS4O8, 2017.4438; found, 2017.4368.
Anal. calcd for C124H212NiS4O8: C, 73.80; H, 10.59; S, 6.36;
found: C, 73.88; H, 10.74; S, 6.29%.
General
All reactions were carried out under an inert atmosphere of
dry N2 using standard Schlenk techniques. 1H and 13C{1H}
NMR spectra were recorded on Varian Mercury 300 spectro-
meters and referenced to the residual proton solvent signals.
Elemental analyses were performed at Atlantic Microlab, Inc.,
Norcross, GA. High-resolution mass spectral analyses were
performed by the Bioanalytical Mass Spectrometry Facility at
Georgia Institute of Technology. POM images were acquired
using an Olympus BX51 microscope. DSC measurements
were performed using a Shimadzu DSC/TGA 50, under an
N2-purged atmosphere, at a scan rate of 10 uC min21. X-Ray
4,49-{4-[5-(3,4,5-Trisdecyloxyphenyl)-[1,3,4]oxadiazol-2-
yl]benzyloxy}benzil (S2). A mixture of 4,49-dihydroxybenzil36
(0.48 g, 2 mmol), 2-(4-bromomethylphenyl)-5-(3,4,5-trisdecyl-
oxyphenyl)-[1,3,4]oxadiazole37 (3.14 g, 4 mmol), and K2CO3
(0.55 g, 4 mmol) in DMF (60 mL) was heated at 100 uC for
3 days. The reaction mixture was allowed to cool down to
room temperature and poured into H2O. The mixture was
extracted with CH2Cl2, and the organic layer was washed with
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 2642–2647 | 2645

View Article Online
H2O and dried over anhydrous MgSO4. The solvent was
removed under reduced pressure. The crude product was
washed with hot hexanes to give a white solid (1.61 g, 48.8%).
1H NMR (300 MHz, CDCl3) d 8.14 (d, J = 8.5 Hz, 4H), 7.95
(d, J = 8.8 Hz, 4H), 7.57 (d, J = 8.2 Hz, 4H), 7.29 (s, 4H), 7.04
(d, J = 8.8 Hz, 4H), 5.21 (s, 4H), 4.06 (t, J = 6.6 Hz, 8H), 4.02
(t, J = 6.6 Hz, 4H), 1.70–1.88 (m, 12H), 1.20–1.54 (m, 84H),
0.86 (t, J = 6.6 Hz, 18H). 13C{1H} NMR (75 MHz, CDCl3) d
192.94, 164.66, 163.79, 163.40, 153.44, 141.28, 139.43, 132.34,
127.65, 127.17, 126.57, 123.78, 118.26, 115.01, 105.35, 73.62,
69.55, 69.37, 32.00, 31.96, 30.40, 29.79, 29.73, 29.70, 29.64,
29.46, 29.41, 29.38, 26.15, 22.77, 14.22. HRMS-MALDI (m/z):
[M+H]+ calcd for C104H151N4O12, 1648.1323; found,
1648.1373. Anal. calcd for C104H150N4O12: C, 75.78; H, 9.17;
N, 3.40; found: C, 75.48; H, 9.13; N, 3.37%.
137.47 (d, 1JCF = 254.2 Hz), 132.33, 126.97, 114.80, 109.14 (tt,
3
2JCF = 18.3 Hz, JCF = 3.4 Hz), 57.53. HRMS-FAB (m/z):
[M+H]+ calcd for C28H13F10O4, 603.06542; found, 603.06447.
Anal. calcd for C28H12F10O4: C, 55.83; H, 2.01; F, 31.54;
found: C, 55.71; H, 1.97; F, 31.77%.
Nickel bis{1,2-di-[4-(2,3,4,5,6-pentafluorobenzyloxy)-phenyl]
ethane-1,2-dithiolene} (3). A mixture of S3 (0.90 g, 1.5 mmol),
phosphorus pentasulfide (1.0 g, 2.3 mmol) in 1,4-dioxane
(20 mL) was refluxed at 110 uC for 4 h under an N2
atmosphere. The reaction mixture was allowed to cool down to
ca. 60 uC and filtered under nitrogen. NiCl2?6H2O (0.178 g,
0.75 mmol) dissolved in H2O (1 mL) was added to the filtrate.
The resulting mixture was refluxed at 110 uC for an additional
2 h. The reaction mixture was allowed to cool down to room
temperature and then poured into ethanol. A grayish-green
solid was collected by filtration and washed with ethanol. The
crude product was purified by column chromatography (silica
gel, hexanes : dichloromethane = 1 : 1) to give a grayish-green
solid. The product was further purified by recrystallization
from an ethyl acetate–isopropanol solvent mixture to give a
grayish-green solid (0.31 g, 31.2%). 1H NMR (300 MHz,
CDCl3) d 7.35 (d, J = 8.8 Hz, 8H), 6.87 (d, J = 8.8 Hz, 8H),
5.11 (s, 8H). 13C{1H} NMR (75 MHz, CDCl3) d 180.10,
Nickel bis[1,2-di(4-{4-[5-(3,4,5-trisdecyloxyphenyl)-
[1,3,4]oxadiazol-2-yl]-benzyloxy}-phenyl)ethane-1,2-dithiolene]
(2). A mixture of S2 (0.99 g, 0.6 mmol), phosphorus
pentasulfide (0.4 g, 0.9 mmol), in 1,4-dioxane (20 mL) was
refluxed at 110 uC for 4 h under an N2 atmosphere. The
reaction mixture was allowed to cool down to ca. 60 uC
and filtered through a sintered glass frit under nitrogen.
NiCl2?6H2O (72 mg, 0.3 mmol) dissolved in H2O (1 mL) was
added to the filtrate. The resulting mixture was refluxed at
110 uC for an additional 2 h. The reaction mixture was allowed
to cool down to room temperature. The reaction mixture was
poured into ethanol. A dark green solid was collected by
filtration and washed with ethanol. The crude product was
purified by column chromatography (silica gel, hexanes : ethyl
1
1
158.47, 145.65 (d, JCF = 250.8 Hz), 141.75 (d, JCF
=
255.4 Hz), 137.46 (d, 1JCF = 254.2 Hz), 135.12, 130.41, 114.67,
2
3
109.70 (tt, JCF = 17.2 Hz, JCF = 3.5 Hz), 57.48. HRMS-
MALDI (m/z): [M]+ calcd for C56H24F20NiO4S4, 1325.9592;
found, 1325.9529. UV (CH2Cl2) lmax/nm (e/M21 cm21): 301
(40 190), 319 (sh, 33 600), 460 (sh, 3130), 612 (1540), 909
(28 610). Anal. calcd for C56H24F20NiO4S4: C, 50.66; H, 1.82;
S, 9.66; found: C, 50.53; H, 1.72; S, 9.46%.
1
acetate = 1 : 1) to give a dark green solid (0.19 g, 18.5%). H
NMR (300 MHz, CDCl3) d 8.13 (d, J = 8.2 Hz, 8H), 7.57 (d,
J = 8.2 Hz, 8H), 7.33 (d, J = 8.8 Hz, 8H), 7.29 (s, 8H), 6.88 (d,
J = 9.1 Hz, 8H), 5.12 (s, 8H), 4.05 (t, J = 6.3 Hz, 16H), 4.02 (t,
J = 6.6 Hz, 8H), 1.70–1.88 (m, 24H), 1.16–1.54 (m, 168H),
0.863 (t, J = 6.6 Hz, 12H), 0.859 (t, J = 6.6 Hz, 24H). 13C{1H}
NMR (75 MHz, CDCl3) d 180.14, 164.60, 163.87, 158.95,
153.41, 141.22, 140.13, 134.62, 130.35, 127.68, 127.03, 123.54,
118.29, 114.63, 105.31, 73.60, 69.35, 31.96, 30.40, 29.79, 29.73,
29.70, 29.64, 29.48, 29.41, 29.38, 26.15, 22.75, 14.21. UV
(CH2Cl2) lmax/nm (e/M21 cm21): 306 (172 650), 632 (1980),
920 (40 530). Anal. calcd for C208H300N8NiO20S4: C, 73.06;
H, 8.84; N, 3.28; S, 3.75; found: C, 72.86; H, 8.71; N, 3.32;
S, 3.78%.
Device fabrication and testing
SCLC samples were prepared by sandwiching melt-processed
films of 1, 2 or 3 between electrodes with different work
functions (Ag or ITO). The active area of each device was
0.05 cm2. For SCLC experiments, the measurements of J–V
characteristics of all devices were performed in air and in the
dark using a Keithley 2400 source meter. Capacitance was
measured using Agilent 16048A test leads connected to an
Agilent 4284A Precision LCR meter, yielding values of 4 pF.
Transistors were fabricated in an N2-filled glovebox by spin-
coating (1000 rpm, 60 s) a solution of 1 in chloroform
(20 mg mL21) onto heavily n-doped Si substrates with 200 nm
of thermally-grown SiO2 as the gate dielectric. Top contact
geometry Ag (100 nm) source and drain electrodes were
evaporated through a shadow mask defining channels with
widths ranging from 200 mm to 2000 mm and lengths ranging
from 25 mm to 200 mm. Depositions were performed at room
temperature using physical vapor deposition at a pressure of
5 6 1028 Torr. The typical layer thickness of 1 was 100 nm.
During testing electrical connections were made with a micro-
probe station contained within a second N2-filled glovebox
and an Agilent E5272A medium power source–monitor unit,
connected to the probe station, was used to perform the
electrical measurements. At no point after deposition of the
organic layer were the transistor devices exposed to air. For
4,49-Bis-(2,3,4,5,6-pentafluorobenzyloxy)benzil (S3). A mix-
ture of 1,2-bis-(4-hydroxyphenyl)-ethane-1,2-dione (1.0 g,
4.1 mmol), 2,3,4,5,6-pentafluorobenzyl bromide (2.16 g,
8.3 mmol), and K2CO3 (5.7 g, 41.3 mmol) in acetone (40 mL)
was refluxed at 65 uC overnight. The reaction mixture was
allowed to cool down to room temperature. Celite was added
to the reaction mixture, and the mixture was filtered and
washed with acetone. The filtrate was evaporated under
reduced pressure. The crude product was purified by column
chromatography (silica gel, dichloromethane : hexanes = 2 : 1)
to give a yellow solid (1.54 g, 61.8%). 1H NMR (300 MHz,
CDCl3) d 7.95 (d, J = 8.8 Hz, 4H), 7.03 (d, J = 8.8 Hz, 4H), 5.18
(s, 4H). 13C{1H} NMR (75 MHz, CDCl3) d 192.82, 162.82,
1
1
145.57 (d, JCF = 254.2 Hz), 141.88 (d, JCF = 255.4 Hz),
2646 | J. Mater. Chem., 2007, 17, 2642–2647
This journal is ß The Royal Society of Chemistry 2007

View Article Online
9 P. K. Ng, X. Gong, S. H. Chan, L. S. M. Lam and W. K. Chan,
Chem.–Eur. J., 2001, 7, 4358.
10 W. K. Chan, P. K. Ng, X. Gong and S. Hou, J. Mater. Chem.,
1999, 9, 2103.
11 X. F. Ren, B. D. Alleyne, P. I. Djurovich, C. Adachi, I. Tsyba,
R. Bau and M. E. Thompson, Inorg. Chem., 2004, 43, 1697.
12 N. B. McKeown, Phthalocyanine Materials: Synthesis, Structure,
and Function, Cambridge University Press, Cambridge, 1998.
13 K. W. Chan, P. K. Ng, X. Gong and S. Hou, Appl. Phys. Lett.,
1999, 75, 3920.
each individual device current–voltage (IDS vs. VDS at multiple,
discrete VGS values) and transfer (IDS vs. VGS at fixed VDS
)
characteristics were measured. Field-effect mobilities were
pffiffiffiffiffiffiffi
calculated in the saturation re´gime by fitting IDS vs. VGS data
to the square law:
CW
2
I
DS~m
ðVGS{VTÞ
(1)
2L
14 H. L. Wong, L. S. M. Lam, K. W. Cheng, K. Y. K. Man,
W. K. Chan, C. Y. Kwong and A. B. Djurisic, Appl. Phys. Lett.,
2004, 84, 2557.
15 S.-I. Noro, H.-C. Chang, T. Takenobu, Y. Murayama, T. Kanbara,
T. Aoyama, T. Sassa, T. Wada, D. Tanaka, S. Kitagawa, Y. Iwasa,
T. Akutagawa and T. Nakamura, J. Am. Chem. Soc., 2005, 127,
10012.
where m is the field-effect mobility, C is the capacitance density of
the gate dielectric (F cm22), VT is the threshold voltage, and W
(width) and L (length) are the dimensions of the semiconductor
channel defined by the source and drain electrodes.
Conclusion
16 T. Tagushi, H. Wada, T. Kambayashi, B. Noda, M. Goto, T. Mori,
K. Ishikawa and H. Takezoe, Chem. Phys. Lett., 2006, 421, 395.
17 T. D. Anthopoulos, S. Setayesh, E. Smits, M. Colle, E. Cantatore,
B. de Boer, P. W. M. Blom and D. M. de Leeuw, Adv. Mater.,
2006, 18, 1900.
18 Z. S. Herman, R. F. Kirchner, G. H. Loew, U. T. Mueller-
Westerhoff, A. Nazal and M. C. Zerner, Inorg. Chem., 1982, 21, 46.
19 U. T. Mueller-Westerhoff, B. Vance and D. I. Yoon, Tetrahedron,
1991, 47, 909.
20 F. Barriere, N. Camire, W. E. Geiger, U. T. Mueller-Westerhoff
and R. Sanders, J. Am. Chem. Soc., 2002, 124, 7262.
21 B. S. Lim, D. V. Fomitchev and R. H. Holm, Inorg. Chem., 2001,
40, 4257.
22 C. Faulmann and P. Cassoux, Prog. Inorg. Chem., 2003, 52, 399.
23 P. Cassoux and L. Valade, in Inorganic Materials, ed. D. W. Bruce
and D. O’Hare, Wiley, Chichester, 2nd edn, 1996.
24 C. Pearson, J. E. Gibson, A. J. Moore, R. M. Bryce and
M. C. Petty, Electron. Lett., 1993, 29, 1377.
25 K. Ohta, H. Hasebe, H. Ema, T. Fujimoto and I. Yamamoto,
J. Chem. Soc., Chem. Commun., 1989, 1610.
26 K. Ohta, H. Hasebe, H. Ema, M. Moriya, T. Fujimoto and
I. Yamamoto, Mol. Cryst. Liq. Cryst., 1991, 208, 21.
27 K. Ohta, H. Hasebe, H. Ema, M. Moriya, T. Fujimoto and
I. Yamamoto, Mol. Cryst. Liq. Cryst., 1991, 208, 33.
28 Solid-state HOMO and LUMO energies were estimated using
In summary, we have found that an electron mobility as high
as 2.8 cm2 V21 s21 can be achieved for the square-planar Ni
bis(dithiolene) complex 1 in an annealed unsymmetrical melt-
processed device, as measured by SCLC experiments in a
direction perpendicular to the substrate. It is also worth noting
that the three Ni bis(dithiolene) complexes we have studied
have very different observed SCLC mobilities despite having
similar molecular electronic properties (as indicated by
electrochemical and optical data); these large differences in
mobility may be correlated with the differing degrees of order
in films of these materials. Field-effect transistors based on
spin-coated
1 show n-channel behavior with saturation
electron mobilities up to 1.3 6 1023 cm2 V21 s21. One factor
that may contribute to the discrepancy between SCLC and
FET mobilities is the different molecular packing present in
the films used for the two types of devices. These results
further indicate the promise of Ni bis(dithiolene) complexes
for organic electronics applications, while underscoring the
importance of controlling solid-state packing in molecular
electronic materials.
+/0
+/0
E1/2
(M) – E1/2
(TPD) = EHOMO (M) – EHOMO (TPD)
0/–
0/–
and E1/2 (M) – E1/2 (Alq3) = ELUMO (M) – ELUMO (Alq3)
respectively (electrochemical potentials in V, orbital energies in
eV), using values of E1/2 (TPD+/0) = 0.26 V,3 E1/2 (Alq30/–) =
Acknowledgements
22.30 V and EHOMO (TPD) = 5.4 eV ,29 ELUMO (Alq3) = 2.3 eV30
.
29 J. D. Anderson, E. M. McDonald, P. A. Lee, M. L. Anderson,
E. L. Ritchie, H. K. Hall, T. Hopkins, E. A. Nash, J. Wang,
A. Padias, S. Thayumanavan, S. Barlow, S. R. Marder, G. Jabbour,
S. Shaheen, B. Kippelen, N. Peyghambarian, R. M. Wightman and
N. R. Armstrong, J. Am. Chem. Soc., 1998, 120, 9646.
30 A. Kahn, N. Koch and W. Y. Gao, J. Polym. Sci., Part B: Polym.
Phys., 2003, 41, 2529.
31 N. F. Mott, D. Gurney, Electronic Processes in Ionic Crystals,
Academic Press, New York, 1970, p. 45.
32 P. N. Murgatroyd, J. Phys. D: Appl. Phys., 1970, 3, 151.
33 A. J. Campbell, D. D. C. Bradley and D. G. Lidzey, J. Appl. Phys.,
1997, 82, 6326.
34 Z. An, J. Yu, S. C. Jones, S. Barlow, S. Yoo, B. Domercq, P. Prins,
L. D. A. Siebbeles, B. Kippelen and S. R. Marder, Adv. Mater.,
2005, 17, 2580.
35 J. Laugier and B. Bochu, LMGP-Suite: Suite of Programs for the
Interpretation of X-ray Experiments, ENSP/Laboratoire des
Mate´riaux et du Ge´nie Physique, BP 46, 38042 Saint Martin
d’He`res, France. Available on the WWW: http://www.inpg.fr/
LMGP or http://www.ccp14.ac.uk/tutorial/lmgp/.
This material is based upon work supported in part by the
STC Program of the National Science Foundation under
Agreement Number DMR-0120967. We also thank the NSF
(CHE-0211419) and ONR (N00014-04-1-0120) for support.
References
1 H. E. Katz, Chem. Mater., 2004, 16, 4748.
2 V. Lemaur, D. A. da Silva Filho, V. Coropceanu, M. Lehmann,
Y. Geerts, J. Piris, M. G. Debije, A. M. van de Craats,
K. Senthilkumar, L. D. A. Siebbeles, J. M. Warman, J.-L. Bre´das
and J. Cornil, J. Am. Chem. Soc., 2004, 126, 3271.
3 R. D. Hreha, C. P. George, A. Haldi, B. Domercq, M. Malagoli,
S. Barlow, J.-L. Bre´das, B. Kippelen and S. R. Marder, Adv. Funct.
Mater., 2003, 13, 967.
4 P. W. Dimmock, J. McGinnis, B. L. Ooi, A. G. Sykes and
A. Geoffrey, Inorg. Chem., 1990, 29, 1085.
5 R. M. Nielson, M. N. Govin, G. E. McManis and M. J. Weaver,
J. Am. Chem. Soc., 1988, 110, 1745.
36 G. Henry and B. H. Smith, J. Am. Chem. Soc., 1948, 70, 2619.
37 Synthesized by analogy with its tris(hexyloxy) analogue (S. Marder,
B. Kaafarani, S. Barlow, B. Kippelen, B. Domercq, Q. Zhang and
T. Kondo, PCT Int. Appl., WO 2005123737, 2005).
6 N. Sutin, M. J. Weaver and E. I. Yee, Inorg. Chem., 1980, 19, 1096.
7 M. S. Chan and A. C. Wahl, J. Phys. Chem., 1982, 86, 126.
8 A. Kokil, I. Shiyanovskaya, K. D. Singer and C. J. Weder, J. Am.
Chem. Soc., 2002, 124, 9978.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 2642–2647 | 2647