Influence of methoxy groups on the properties of 1,1-bis(4-aminophenyl) cyclohexane based arylamines: Experimental and theoretical approach

Article abstract of DOI:10.1039/c2jm14387a

Three new isomeric cyclohexylidene linked triphenylamines containing methoxy groups in different positions were synthesized via Ullmann coupling from 1,1-bis(4-aminophenyl)cyclohexane and respective aryl iodides. Thermal behaviour, optical and photoelectr

Full text of DOI:10.1039/c2jm14387a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 3015
www.rsc.org/materials
PAPER
Influence of methoxy groups on the properties of 1,1-bis(4-aminophenyl)
cyclohexane based arylamines: experimental and theoretical approach†
a
Jonas Keruckas, Ramunas Lygaitis, Jurate Simokaitiene, Juozas Vidas Grazulevicius,
a
a
a
*
b
Vygintas Jankauskas and Gjergji Sini
c
*
Received 5th September 2011, Accepted 5th December 2011
DOI: 10.1039/c2jm14387a
Three new isomeric cyclohexylidene linked triphenylamines containing methoxy groups in different
positions were synthesized via Ullmann coupling from 1,1-bis(4-aminophenyl)cyclohexane and
respective aryl iodides. Thermal behaviour, optical and photoelectrical properties of the obtained
materials were investigated. Calculations based on the density functional methods (DFT) were also
carried out in order to better understand structure–property relationships. Methoxy-substituted
derivatives of 1,1-bis(4-aminophenyl)cyclohexane show lower ionization potentials and higher hole
drift mobilities than the corresponding derivative having no methoxy groups. The ionization potentials
of the solid samples of the methoxy-substituted compounds established by the electron photoemission
technique are in the range of 5.34–5.55 eV. Hole-drift mobility values of the amorphous layers of the
methoxy-substituted materials established by the time-of-flight technique range from 4.0 ꢀ 10ꢁ4 to
1.2 ꢀ 10ꢁ3 cm2 Vꢁ1 sꢁ1 at the electric field of 106 V cmꢁ1. The highest hole mobilities were observed for
the para-substituted derivative. Theoretical results suggest that the hole mobility in the bulk materials is
driven by the electronic coupling parameter whilst the reorganization energy parameter predicts the
wrong mobility trend. The role of the methoxy groups is found to be related to the well-known
mesomeric (p-donor) effect and the possibility to establish C–H/p(Ph) and C–H/X (X ¼ O, N)
hydrogen bonds. The effect of these properties on the enhanced Stokes shifts of the para-methoxy-
substituted compound, on the decrease of the ionization potentials for the methoxy substituted
compounds as compared to the non-substituted one, and on the enhanced possibility to establish
considerable electronic couplings between adjacent molecules is discussed in detail.
ionization potentials can be varied in a rather wide range, and
their devices show stable performance under ambient conditions.
In some fields of application, such as dye sensitized solar cells,
hole-transporting materials with low ionization potentials are
required. Triphenylamine (TPA) based compounds have attrac-
ted much interest in this respect and the role of different
substituents in the ionization potentials and other parameters
has been considered.5–9 It is known that introduction of methoxy
groups into the structures of organic p-type semiconductors
leads to the decrease of their ionization potentials,5 with stronger
influence found in the case of para-methoxy substituted TPA
compounds as compared to the meta-substituted ones.6 As for
the influence of these substitutions on the hole mobilities, lower
mobility has been measured for instance in the case of p-methoxy
and p-butoxy substituted N,N0-bis(m-tolyl)-N,N0-diphenyl-1,10-
biphenyl-4,40-diamine (TPD) as compared to the non-substituted
TPD7 which has been partly explained by the change in dipole
moment upon substitution. Due to this kind of substitutions, the
efficiencies of different devices may be strongly affected. The
hole-injection barrier differences for instance, which are strongly
related to the highest occupied molecular orbital (HOMO)
1. Introduction
Arylamines represent a class of organic charge-transporting
materials widely studied and used as the components of different
optoelectronic devices including electrophotographic photore-
ceptors, organic light emitting diodes, solar cells, etc.1–4 They
exhibit quite sufficient charge transport properties, their
aDepartment of Organic Technology, Kaunas University of Technology,
Radvilenu plentas 19, LT-50254 Kaunas, Lithuania. E-mail: juozas.
grazulevicius@ktu.lt
bDepartment of Solid State Electronics, Vilnius University, Sauletekio aleja
9, LT-10222 Vilnius, Lithuania
cLaboratoire de Physicochimie des Polymeꢀres et des Interfaces, EA 2528
Universiteꢁ de Cergy-Pontoise, 5 mail Gay-Lussac, 95031 Cergy-Pontoise
Cedex, France. E-mail: gjergji.sini@u-cergy.fr
† Electronic supplementary information (ESI) available: Values of Dd
parameter and DE(H ꢁ Hꢁ1); excitation energy values of the first ten
excited states for compounds 1–4 along with the corresponding
oscillator strengths; optimized geometries of compounds 2–4; UV-VIS
absorption spectra from the TDDFT calculations; sketches of some
relevant orbitals for compound 4; details on the estimation of kHT
values. See DOI: 10.1039/c2jm14387a
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 3015–3027 | 3015

View Article Online
energy, could dominate the efficiencies of devices using this kind
of materials as has been shown in the case of some OLED
applications.8
2. Experimental
2.1 Instrumentation
It is interesting to note that while these studies point to the
effect of the para-methoxy substitutions on some charge-trans-
port properties, there is a lack of systematic studies concerning
the role of the methoxy group position in the optical, electronic,
and hole-transporting properties of the TPA based compounds.
A deeper insight on the mechanisms by which the methoxy group
influences these properties may be very helpful in the design of
new hole-transporting materials.
NMR spectra were recorded using a Varian Unity Inova (300
MHz) apparatus. IR spectra were obtained on a Perkin-Elmer
Spectrum GX II FT-IR System spectrometer. Mass spectra (MS)
were obtained using a Waters ZQ 2000 system. Elemental anal-
yses were performed with an Exeter Analytical CE-440
Elemental Analyzer. UV spectra were recorded on a Perkin
Elmer Lambda 35 spectrophotometer. Photoluminescence (PL)
spectra were recorded using a Hitachi MPF-4 spectrophotom-
eter. Differential scanning calorimetry (DSC) measurements
were carried out using a DSC Q100 calorimeter with a heating/
In order to understand experimental results and to predict
strategies for designing new charge transporting materials,
theoretical approaches on the molecular level have been exten-
sively employed. Different strategies have been used for the
charge-mobility calculations.10–12 In the case of amorphous
materials one can consider that the elementary step of the charge
transport is the charge transfer between two adjacent molecules
in which case the determination of hole-transfer rate-constants
(kHT, eqn (1)) based on Marcus13–16 theory can give very helpful
information.
cooling rate of 10 C minꢁ1 under nitrogen. Thermogravimetric
ꢂ
analysis (TGA) was performed on a Mettler TGA/SDTA851e/
LF/1100 apparatus with a heating rate of 20 ꢂC minꢁ1 under
nitrogen. The ionization potentials (Ip) of the solid samples were
measured by electron photoemission in the air method as
described before.18 The measurement method was, in principle,
similar to that described by Miyamoto et al.19 Hole drift mobil-
ities were measured by a xerographic time-of-flight (XTOF)
method.20,21 The samples for the charge carrier mobility
measurements were prepared by casting tetrahydrofuran (THF)
solutions of compounds or their mixtures with bisphenol-Z
polycarbonate (PC-Z) on polyester films with aluminium layer.22
The thickness of the charge-transporting layers varied in the
range of 2–8 mm.
"
#
2
4p
1
ðDGꢂ þ lÞ
4kBlT
2
t exp
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
kHT
¼
ꢁ
(1)
h
4plkBT
In this equation t is the electronic coupling between two
adjacent molecules, DGꢂ is the CT reaction free energy (which is
zero in the case of hole transfer between identical molecules) and
l is the reorganization energy. This last parameter is the sum of
two terms: (i) ls, representing the contribution from the medium
polarization energy and (ii) li, representing the energetic effort
due to the intramolecular geometric relaxations upon CT
between two adjacent molecules.
2.2 Materials
1,1-Bis(4-aminophenyl)cyclohexane was purchased from Tokyo
Chemical Industry. Iodobenzene, 2-, 3- and 4-iodoanisoles, o-
dichlorobenzene, 18-crown-6, copper powder and potassium
carbonate were obtained either from Sigma-Aldrich or Fluka.
All the materials were used as received. Compounds 1–4 were
synthesized by the following general procedure. 1,1-Bis(4-ami-
nophenyl)cyclohexane (0.8 g, 3 mmol), aryl iodide (18 mmol),
either iodobenzene (3.7 g) or iodoanisole (4.2 g), and 18-crown-6
(0.16 g, 0.6 mmol) were dissolved in 10 ml of o-dichlorobenzene
in a two-necked round-bottom flask equipped with a magnetic
stirrer and under nitrogen blanket. When the mixture was heated
In order to obtain high kHT values the electronic coupling
parameter (t) should be maximized whereas the reorganization
energy (l ¼ ls + li) should be minimized. While the knowl-
edge of all these parameters is necessary for the calculation of
kHT, still other parameters need to be known in order to
calculate charge mobilities. However, each of these parameters
considered individually can be helpful in designing new
materials. The intramolecular reorganization energy param-
eter, for instance, has been frequently used to deduce trends in
ꢂ
above 150 C dry potassium carbonate powder (1 g, 7.5 mmol)
a
series of similar compounds, which is based on the
and copper bronze (0.38 g, 6 mmol) were added. The mixture was
kept under reflux (ca. 180 ꢂC) controlling by TLC. The reaction
was stopped after ca. 20 hours, the mixture was cooled down,
filtered and o-dichlorobenzene was removed by distillation. The
products were further purified by silica gel column chromatog-
raphy and crystallized from the methanol and tetrahydrofuran
(THF) mixture. Compound 1 was crystallized from the eluent
mixture.
assumption that, in this case, the variation of the electronic
couplings should be smaller than the variation of the reorga-
nization energies. While this might hold true in some cases, in
the present study we show that the application of this criterion
is subtle.
The aim of this work is the investigation of the influence of the
methoxy group position on ionization potentials, charge-trans-
porting and other properties of 1,1-bis(4-aminophenyl)cyclo-
hexane based arylamines. With this aim we have synthesized and
characterized three new isomeric 1,1-bis(4-aminophenyl)cyclo-
hexane linked triphenylamines containing methoxy groups in
different positions. Density functional calculations (DFT)17 were
also carried out in order to obtain more insight on the mecha-
nisms by which the presence and the positioning of the methoxy
groups influence the electronic, optical and hole-transporting
properties.
1,1-Bis[4-(N,N-diphenyl)aminophenyl] cyclohexane (1). Pale
1
grey crystals. Yield 1.09 g (64%). Mp: 171–173 ꢂC. H NMR
(CDCl3, 300 MHz) ppm: 1.46–1.65 (m, 6H, CH2 cyclo-
hexylidene), 2.16–2.28 (m, 4H, CH2 cyclohexylidene), 6.94–7.04
(m, 8H, phenyl), 7.04–7.10 (m, 4H, phenyl), 7.07 (d, 4H, p-phe-
nylene, J ¼ 8.8 Hz), 7.13 (d, 4H, p-phenylene, J ¼ 8.8 Hz), 7.18–
7.28 (m, 8H, phenyl). 13C NMR (CDCl3, 300 MHz) ppm: 22.9,
26.4, 37.3, 45.5, 122.4, 123.5, 124.1, 127.9, 129.1, 142.9, 144.9,
3016 | J. Mater. Chem., 2012, 22, 3015–3027
This journal is ª The Royal Society of Chemistry 2012

View Article Online
147.9. FT-IR (KBr) cmꢁ1: 3034 (n C–H aromatic), 2963, 2926,
2855 (n C–H aliphatic), 1592, 1507, 1493, 1465 (n C]C
aromatic), 1327 (n C–N aromatic), 862, 803, 695 (g C–H
aromatic). MS (APCI+, 20 V) m/z 571.5 ([M + H]+). Elemental
analysis for C42H38N2: calc. C 88.38%; H 6.71%; N 4.91%; found:
C 88.34%; H 6.69%; N 4.98%.
the molecules were optimized without symmetry constraints by
using the 6-31G** basis set, followed by frequency calculations
to assure that real minima were obtained. The geometry opti-
mizations of the cationic radical species were performed at the
restricted open shell level.
The spectroscopic properties of the molecules were calculated
by means of the time dependent density functional theory
(TDDFT) method26–30 with the 6-31+G* basis set. Up to 20
excited states were calculated and the theoretical absorption
bands were obtained by considering a band half-width of 0.33 eV
at half-height (Gaussview 5 software).
1,1-Bis{4-[N,N-bis(2-methoxyphenyl)]aminophenyl}cyclohexane (2).
ꢂ
1
Yellow crystals. Yield 1.04 g (50%). Mp: 179–180 C. H NMR
(CDCl3, 300 MHz) ppm: 1.36–1.57 (m, 6H, CH2 cyclo-
hexylidene), 2.04–2.20 (m, 4H, CH2 cyclohexylidene), 3.61 (s,
12H, OCH3), 6.55 (d, 4H, p-phenylene, J ¼ 8.8 Hz), 6.83–6.92
(m, 8H, o-phenylene), 6.97 (d, 4H, p-phenylene, J ¼ 8.8 Hz),
7.06–7.17 (m, 8H, o-phenylene). 13C NMR (CDCl3, 300 MHz)
ppm: 23.0, 26.6, 37.3, 44.9, 55.7, 112.8, 117.2, 121.0, 125.4, 127.1,
128.3, 135.9, 140.0, 145.3, 155.2. FT-IR (KBr) cmꢁ1: 3059, 3029,
2991 (n C–H aromatic), 2936, 2849, 2834 (n C–H aliphatic), 1612,
1591, 1500 (n C]C aromatic), 1326 (n C–N aromatic), 1271,
1237, 1024 (n C–O–C), 820, 743, 697, 621 (g C–H aromatic). MS
(APCI+, 20 V) m/z 691.3 ([M + H]+). Elemental analysis for
C46H46N2O4: calc. C 79.97%; H 6.71%; N 4.05%; found C
79.88%; H 6.69%; N 4.08%.
The internal reorganization energies (li) were calculated with
the 6-31G** basis set by means of the following equation:31
ꢀ
ꢁ
ꢀ
ꢁ
þ
þ
li ¼ l1i þ li2 ¼ EMGꢁM ꢁ EMGꢁM
þ
þ
EMGþꢁM ꢁ EMGþꢁM
(2)
In this equation the quantity EMGꢁM for instance corresponds
to the energy of the neutral molecule (M) in the geometry of the
cationic species (M+).
The calculation of the electronic couplings is based on the
following approximations: (i) despite the irregular packing
between adjacent molecules in the amorphous materials, dimers
of different geometries can be established which are supposed to
adopt local minima geometries. (ii) The hole is localized in only
one molecule. (iii) Electronic coupling between the HOMO
orbitals of adjacent monomers, also known as the ‘‘direct
calculation’’ or ‘‘two state model,’’32–35 is considered instead of
the coupling between states. Detailed descriptions on different
methods employed for the calculation of the transfer integrals
can be found elsewhere.31–35 The simplest but the more limited
method is based on the Koopmans’ theorem, in which the
transfer integrals are determined as half of the energy splitting
between the HOMO-1 (LUMO-1) and HOMO (LUMO) orbitals
in the dimer. As this method is applicable only if the dimer
possesses a perfect symmetry, the ‘‘direct calculation’’ of the
transfer integrals in the frame of a ‘‘two state’’ model is
commonly employed. Some more details on this method are
given in the ESI†, but we emphasize here that the HOMO
(LUMO) orbitals of each monomer are generally considered,
with the strict condition that the dimer HOMO (LUMO)
contains contributions only from the monomer HOMO
(LUMO) orbitals. As shown later in this study, this condition is
satisfied for all compounds 1–4 which justifies the use of this
method in our study. Finally, a general method evaluating the
electronic coupling between electronic states (instead of frontier
orbitals) has been proposed.36 However, in our study the two-
state model making use of the frontier orbitals for the calculation
of the transfer integrals has been employed, which is commonly
used and has proved to give correct results.
1,1-Bis{4-[N,N-bis(3-methoxyphenyl)]aminophenyl}cyclohexane (3).
1
Pale yellow crystals. Yield 0.77 g (37%). Mp: 121–123 ꢂC. H
NMR (CDCl3, 300 MHz) ppm: 1.41–1.63 (m, 6H, CH2 cyclo-
hexylidene), 2.14–2.26 (m, 4H, CH2 cyclohexylidene), 3.70 (s,
12H, OCH3), 6.51–6.58 (m, 4H, m-phenylene), 6.62 (t, 4H,
m-phenylene, J ¼ 2.2 Hz), 6.63–6.69 (m, 4H, m-phenylene), 6.99
(d, 4H, p-phenylene, J ¼ 8.8 Hz), 7.13 (d, 4H, p-phenylene, J ¼
8.8 Hz), 7.13 (t, 4H, m-phenylene, J ¼ 8.1 Hz). 13C NMR
(CDCl3, 300 MHz) ppm: 22.9, 26.4, 37.2, 45.5, 55.2, 107.8, 109.9,
116.6, 124.1, 127.9, 129.7, 143.3, 144.6, 149.0, 160.3. FT-IR
(KBr) cmꢁ1: 3031, 2997 (n C–H aromatic), 2934, 2857, 2832 (n C–
H aliphatic), 1596, 1506, 1487 (n C]C aromatic), 1318 (n C–N
aromatic), 1208, 1043 (n C–O–C), 826, 769, 691 (g C–H
aromatic). MS (APCI+, 20 V) m/z 691.4 ([M + H]+). Elemental
analysis for C46H46N2O4: calc. C 79.97%; H 6.71%; N 4.05%;
found C 80.00%; H 6.77%; N 4.11%.
1,1-Bis{4-[N,N-bis(4-methoxyphenyl)]aminophenyl}cyclohexane (4).
1
Pale yellow crystals. Yield 0.93 g (45%). Mp: 162–163 ꢂC. H
NMR (CDCl3, 300 MHz) ppm: 1.41–1.62 (m, 6H, CH2 cyclo-
hexylidene), 2.12–2.22 (m, 4H, CH2 cyclohexylidene), 3.78 (s,
12H, OCH3), 6.80 (d, 8H, p-phenylene, J ¼ 9.2 Hz), 6.82 (d, 4H,
p-phenylene, J ¼ 8.8 Hz), 7.03 (d, 8H, p-phenylene, J ¼ 9.2 Hz),
7.05 (d, 4H, p-phenylene, J ¼ 8.8 Hz). 13C NMR (CDCl3, 300
MHz) ppm: 22.9, 26.5, 37.2, 55.5, 114.5, 120.3, 120.4, 126.3,
126.3, 127.6. FT-IR (KBr) cmꢁ1: 3037, 2996 (n C–H aromatic),
2933, 2857, 2832 (n C–H aliphatic), 1604, 1504 (n C]C
aromatic), 1318 (n C–N), 1239, 1036 (n C–O–C), 826, 729 (g C–H
aromatic). MS (APCI+, 20 V) m/z 691.4 ([M + H]+). Elemental
analysis for C46H46N2O4: calc. C 79.97%; H 6.71%; N 4.05%;
found C 80.03%; H 6.67%; N 4.09%.
In order to correctly describe the p-stacking interactions, the
geometries of the dimers formed by two identical molecules have
been optimized by employing the uB97X-D functional at the 6-
31G** basis set.37 Previous studies have shown that this func-
tional provides good results on the description of weak interac-
tions.38,39 The electronic couplings between the HOMO orbitals
of two adjacent molecules have been calculated at the same level
(uB97X-D/6-31G**) according to the approach described by
Valeev et al.34 with the corresponding matrix elements evaluated
with Gaussian 09 (ESI†). Note that the coupling values depend
2.3 Computational methodology
DFT17 calculations employing the B3LYP23,24 functional were
performed with the Gaussian 09 program.25 The geometries of all
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 3015–3027 | 3017

View Article Online
Table 1 Thermal characteristics of compounds 1–4
on the functional used and generally increase with the increasing
percentage of Hartree–Fock exchange in the functional.38 The
transfer integrals for 1–4 calculated by using two different
functionals are thus expected to change by a roughly constant
factor, but the trends obtained by both functionals should be the
same.
a
a
Compound
Tm/ꢂC
Tm /ꢂC
Tg /ꢂC
TD/ꢂC
1
2
3
4
171
179
121
162
172
183
123
163
68
67
48
63
416
421
454
435
The interaction energies of some optimized dimers were
calculated with respect to the isolated monomers with the 6-311+
+G(3df,3pd) basis set and were corrected for the basis set
superposition error (BSSE) by the counterpoise correction
method of Boys and Bernardi.40 The zero point energy (ZPE)
correction has been also taken into account by using the value
obtained at the 6-31G** basis set.
a
Established by DSC.
transformed into the glassy state by fast cooling from the melt it
readily crystallizes from solutions or from films prepared by spin
coating. Thus incorporation of OCH3 groups improves the
morphological stability of the glasses without significant reduc-
tion of Tg, which is 68 ꢂC, 67 ꢂC and 63 ꢂC for 1, 2 and 4
respectively. Tg of meta-substituted compound 3 was found to be
Finally, the vertical ionization potentials (Ip) were calculated
at the B3LYP/6-31G** level by the energy difference between
neutral and cation radical species at the neutral state geometry.
ꢂ
considerably lower (48 C).
ꢂ
3. Results and discussion
TGA performed under nitrogen with a heating rate of 20 C
minꢁ1 revealed high thermal stability of the synthesized
compounds. Their thermal degradation started above 400 ꢂC.
Methoxy-substituted derivatives showed slightly higher 5% mass
loss temperatures than compound 1. The highest thermal
stability was observed for meta-substituted derivative 3. Its 5%
3.1 Synthesis
Compounds 1–4 were obtained from primary aromatic diamine,
1,1-bis(4-aminophenyl)cyclohexane, and iodobenzene or
respective iodoanisoles in one-step synthesis via the modified
Ullmann method41 as shown in Scheme 1. The small amounts of
biphenyl derivatives were also obtained due to dimerization of
iodoarenes. The chemical structure of the compounds was
ꢂ
mass loss was observed at 454 C.
3.3 Geometries
1
proved by H and 13C NMR, FT-IR, mass spectrometry and
The geometries of compounds 1–4 have been determined theo-
retically as described in the Computational methodology. Some
typical bond-lengths and dihedral angles are given in Fig. S1
(ESI†). Two possible geometries for compound 1 are shown in
Fig. 2. The two geminal phenyl groups bonded to the cyclo-
hexane moiety can adopt a face-to-face relative geometry
(structure 1a) or can be positioned almost orthogonal to each
other (structure 1b, which mostly resembles to a transition state
structure). Structure 1b is roughly 0.5 kcal molꢁ1 higher in energy
than 1a, suggesting fluxional geometries for this compound. As
for the cationic state, structure 1b is unstable and only structure
1a is obtained.
elemental analysis. The obtained data were in good agreement
with the designed structures. The obtained materials are well-
soluble in common organic solvents such as acetone, THF,
chloroform, toluene, but almost insoluble in methanol and
hexane.
3.2 Thermal characterization
Thermal behaviour of the materials was explored using DSC and
TGA methods. Melting points (Tm), glass transition tempera-
tures (Tg) and 5% mass loss temperatures (TD) were established.
The data are summarized in Table 1. Thermal behaviour of 2
during DSC heating–cooling–heating scans is shown in Fig. 1.
Such behaviour is typical for all other samples. In the first DSC
heating scan sharp endothermic peaks of melting appeared
showing the crystalline nature of the materials. During the
cooling and the second heating scans only glass transitions were
observed for all four samples. Although compound 1 can be
Geometries equivalent to 1a and 1b have been obtained for the
methoxy substituted neutral compounds (2–4, Fig. S1, ESI†)
while only structures similar to 1a have been obtained for the
cationic species. Due to multiple relative orientations of the
methoxy groups, compounds 2–4 can adopt different conformer
geometries with energy differences less than 0.5 kcal molꢁ1
Scheme 1 Synthesis of compounds 1–4.
3018 | J. Mater. Chem., 2012, 22, 3015–3027
This journal is ª The Royal Society of Chemistry 2012

View Article Online
around the N atom. In the case of compound 4, the Ph–‘‘N-
plane’’ dihedral angles are larger for the para-substituted Ph
rings (ꢃ45ꢂ) than for the unsubstituted ones (ꢃ36 to 37ꢂ).
Similarly, the N–C bond lengths are not equivalent for this
ꢂ
ꢂ
compound (1.417 A for the non-substituted ring and 1.423 A
for the substituted ones). In the case of compound 2 (Fig. S1,
ESI†), the methoxy ortho-substitutions induce important
repulsions with the adjacent Ph rings, resulting in
Ph–‘‘N-plane’’ dihedral angles ranging from ꢃ24 to 26ꢂ for the
non-substituted Ph rings to up to ꢃ60 to 63ꢂ for the substituted
ones.
The changes in the Ph–‘‘N-plane’’ dihedral angles upon
oxidation are more important for 2 (sum of all changes ꢃ36ꢂ)
than for 1 (ꢃ28ꢂ), 3 (ꢃ26ꢂ), and 4 (ꢃ10ꢂ). It can be expected that
upon oxidation this difference affects the efficiency of the p-
conjugation between Ph rings more importantly in the case of 2
than for 1, 3 and 4. Finally, from the geometrical parameters
presented in Fig. S1†, it can be deduced that in both states
(neutral and radical cation), the Ph groups directly bonded to the
cyclohexylidene bridge possess more quinoidal character than
the external Ph groups (see also the values of Dd parameter,
Table S1, ESI†).
Fig. 1 DSC of ortho-substituted compound 2.
Fig. 2 Two geometries for compound 1 obtained by gas phase calcu-
lations at the B3LYP/6-31G** level.
between them. It can be assumed consequently that for each
compound (1–4) a mixture of different conformers of very similar
energetical stability might be obtained. In the following only the
conformers of the lowest energy are considered.
In good agreement with previous studies42–44 planar geome-
tries around N atoms are found for all compounds (referred
hereafter as ‘‘N-plane’’) which is due to the conjugation of the
N lone-pair with the p(Ph) systems. The dihedral angles
between the Ph rings and the ‘‘N-plane’’ are almost the same
for
1
and
3
(ꢃ40 to 42ꢂ, absolute values). The N–C
ꢂ
bond lengths (ꢃ1.421 A) are also equivalent, meaning that the
meta-substitutions do not noticeably modify the local geometry
Fig. 3 UV (3a) and PL (3b) spectra of dilute tetrahydrofuran solutions
of compounds 1–4.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 3015–3027 | 3019

View Article Online
3.4 Optical properties
The interpretation of the fluorescence spectra is not straight-
forward. All the excited states involved in the absorption bands
(Fig. S2†) will relax and participate in the observed Stokes shifts.
However, the analysis of the strongest transition for each
compound might shed light on their PL spectra.
UV spectra of the dilute (10ꢁ5 M) THF solutions of compounds
1–4 are shown in Fig. 3a. The wavelengths of absorption maxima
(lUV ) and absorption edges are summarized in Table 2. The
max
spectra of the compounds show wide and intensive absorption
bands with the maxima in the range of 302–306 nm due to p–p*
transitions. UV spectra of the dilute solutions of 1–4 are very
similar to that of dilute solution of triphenylamine (TPA) in
chloroform, which shows maximum at 300 nm.45 The absorption
edges of methoxy-substituted derivatives 2–4 are slightly red
shifted with respect to compound 1.
In the case of compound 4, the strongest transition (S0 / S3)
corresponds principally to HOMO / LUMO + 2 electronic
transition (orbital pictograms shown in Fig. 4i and j). One can
observe that during the S0 / S3 transition some charge amount
is transferred from the external phenyl groups of the TPA
moieties toward the central part of the molecule (containing
principally two phenyl rings bridged by the cyclohexylidene
unit). It can be assumed consequently that the relaxation of this
excited state involves important geometrical reorganizations in
all the molecular backbone, giving rise to considerable Stokes
shifts. The same effect should be produced for the other
compounds when considering the same ground-to-excited state
transition.
PL spectra of the dilute solutions of compounds 1–4 are shown
in Fig. 3b. PL maxima (lPL ) of compounds 1–4 are in the
max
region of 356–389 nm with the Stokes shifts ranging from 50 to
86 nm. The largest Stokes shifts are observed for compound 4
which emits in the violet light region (389 nm). These observa-
tions show that the position of OCH3 group noticeably influences
the fluorescent properties of methoxyphenyl-substituted 1,1-bis
(4-aminophenyl)cyclohexane derivatives.
In order to get some insight on the differences between 1 and 4
we focus on the role of the methoxy groups. The HOMO orbital
of compound 4 contains non-negligible anti-bonding contribu-
tions from the C(Ph)–O bonds (Fig. 4i) which are considerably
reduced in the LUMO + 2 orbital (Fig. 4j). These contributions
are absent in the case of compound 1 but also in the case of meta-
substituted compound 3 (Fig. 4e). It can be then expected that
during this electronic transition the C–O bonds will be much
more affected by geometrical relaxations in the case of
compound 4 than in compound 3 (and 1) which might be one of
the reasons for the strongest Stokes shift observed for para-
substituted compound 4. This effect should be less pronounced
for compound 2 due to the reduced contribution of the ortho-
methoxy groups in the HOMO orbital as compared to 4 sug-
gesting consequently smaller relaxations upon excitation (Fig. 4c
and d).
In order to get more insight on the nature of the absorption
bands, TDDFT calculations were performed. The UV-VIS
spectra for compounds 1–4 (Fig. S2, ESI†) indicate that lUV of
max
each absorption band results as a combination of electronic
transitions toward several excited states (for the excitation
energies and the oscillator strengths of the first ten excited states
see Table S2, ESI†). It can be observed that the band maxima of 1
and 4 do not correspond exactly to the transition of the largest
oscillator strength but mostly to a group of higher energy tran-
sitions of smaller oscillator strengths which are very close in
energy. In the case of 4 for instance, the sum of the oscillator
strengths corresponding to the transitions S0 / S7,8,11 (separated
by ꢃ0.1 eV, Table S2†) is larger than the oscillator strength of the
strongest one (S0 / S3) thus providing the major contribution to
the absorption band maximum.
The same observations could be done for the transitions of
lower oscillator strength (S0 / S7 and S0 / S8 for compound 4,
Table S2†) which involve HOMO-1, HOMO, LUMO + 3,
LUMO + 4, and LUMO + 5 orbitals (Fig. S3†) suggesting
consequently that the contribution of the C–O bonds in the
Stokes shifts works in the same sense for all the relevant
transitions.
The theoretical lUV values of each absorption band were
max
deduced from Fig. S2† and are shown in Table 2. While differ-
ences of ꢃ20 nm roughly can be observed between the calculated
and the experimental values, the negligible spread of the exper-
imental lUV values for compounds 1–4 (3 nm) is very well
max
reproduced by the theoretical results (5 nm, Table 2).
It is worth remembering that due to the presence of multiple
quasi-isoenergetic conformers for each compound (vide supra)
the absorption spectra should contain similar contributions of
wavelengths slightly different to those shown in Table 2. In
Fig. S2 and Table S2† only one absorption spectrum for each
compound is shown corresponding to the conformer of the
lowest energy.
Admittedly, this analysis is not complete but points to one
possible effect of the methoxy groups on the increasing Stokes
shifts in the order 3 < 2 < 4. The smaller Stokes shift for 3 as
compared to 1 seems to ask for more quantitative analysis and is
out of the scope of this work.
3.5 Ionization potentials
The ionization potentials (Ip) of the solid amorphous samples of
the materials are established from their electron photoemission
spectra which are presented in Fig. 5. Ip values are summarized in
Table 3. The introduction of OCH3 groups leads to the decrease
of Ip: methoxy-substituted compounds 2–4 show lower Ip than
compound 1. The Ip values also depend on the position of OCH3
groups. The lowest Ip is observed for para-substituted compound
4. Ortho-Substituted derivative 2 shows slightly higher Ip while
meta-substituted compound 3 shows the highest Ip among
methoxy-substituted derivatives.
Table 2 UV and PL spectral data for the dilute solutions of compounds
1–4. In parentheses are shown the corresponding theoretical values
Absorption
edge/nm
Compound
lUV /nm
lPL /nm
Stokes shift/nm
max
max
1
2
3
4
303 (322)
302 (327)
306 (327)
303 (325)
363
368
356
389
335
340
345
340
60
66
50
86
3020 | J. Mater. Chem., 2012, 22, 3015–3027
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Fig. 4 Pictograms of some relevant molecular orbitals involved in the strongest electronic transition for compounds 1–4. Atoms in red, grey, blue, and
white colors are oxygen, carbon, nitrogen, and hydrogen respectively. The encircled parts highlight the contribution of the C(Ph)–O(CH3) bonds.
The theoretical vertical Ip values are presented in Table 3
and show the same trend as the experimental values. The
differences between the theoretical and experimental
values are in part due to the lack of the medium effects in
calculations but also to the limited level of calculations
employed.
The influence of the methoxy groups on the Ip values can be
understood in the frame of the Koopmans’ theorem relating for
instance the Ip value to the energy of the HOMO orbital. The
contribution of the methoxy groups to the HOMO energy of
compounds 2–4 is principally related to the p-conjugation of
the oxygen p orbitals (p-donor effect) with the HOMO of
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 3015–3027 | 3021

View Article Online
3.6 Charge mobility
The solid amorphous layers of compounds 2–4 used for charge
mobility measurements were homogeneous, transparent and
stable. The morphological stability of the glassy layer of 1 was
not sufficient due to its crystallization therefore its molecular
mixture with inert polymer bisphenol-Z polycarbonate (PC-Z)
(mass ratio of 1 : 1) was used for the time-of-flight measure-
ments. In order to estimate the influence of OCH3 groups on
charge-transporting properties the layers of molecular mixtures
of compounds 2–4 with PC-Z were also prepared and studied.
Fig. 6 shows electric field dependence on hole mobilities for
amorphous layers of 2–4 and molecular mixtures of 1–4 with PC-
Z. The dependence of the square root of the applied electric field
over the logarithm of mobility was found to be linear as expected
for amorphous organic charge-transporting materials.4
The comparison of hole drift mobilities of PC-Z molecularly
doped with the different derivatives of 1,1-bis(4-aminophenyl)
cyclohexane 1–4 shows that methoxy-substituted derivatives 2–4
exhibit superior charge-transporting properties relative to those
of non-substituted derivative 1. The best charge-transporting
properties among methoxy-substituted derivatives (2–4) were
observed for para-substituted compound 4. Hole mobility in the
Fig. 5 Electron photoemission spectra of the solid samples of 1–4.
compound 1 (Fig. 4c, e, and i): the anti-bonding p(OCH3)–p
(Ph) interaction increases the HOMO energy of compounds 2,
3 and 4 as compared to compound 1 which, according to
Koopmans’ theorem, corresponds to decreasing Ip values in the
same order.46
This effect is more important for compounds 2 and 4 as
compared to compound 3. This can be seen when comparing
Fig. 4e and i; the contribution of the oxygen p orbital is much
more important in the case of the para-substituted compound
than for the meta-substituted compound (Fig. 4i and e). The
case of ortho-substituted compound 2 is intermediate, probably
due to the reduced contribution in the HOMO orbital from
the ortho-substituted Ph rings (Fig. 4c). The increase of the
HOMO levels and the decrease of the Ip values should be
consequently more important for the para-substituted
compound (and less important for the ortho one) than for the
meta-substituted compound where the inductive effect remains
dominant.
We finally remember that in some applications, such as dye
sensitized solar cells, hole-transporting materials with low
ionization potentials are required. The HOMO energies for
compounds 1–4 increase in the order 1 < 3 < 2 < 4
suggesting that based only on this factor, the efficiencies
of devices using these materials should also increase in the
order 1 < 3 < 2 < 4.47
Fig. 6 Electric field dependencies on hole mobilities for amorphous
layers of 2–4 and for molecular mixtures of 1–4 with bisphenol-Z poly-
carbonate (PC-Z).
Table 3 Electronic and hole-transport parameters for compounds 1-4
f
mHT /10ꢁ3
cm2 Vꢁ1 sꢁ1
td/meV
kHT /1011 sꢁ1
e
Compound
Ipa/eV
EHOMOb/eV
lic/eV
1
2
3
4
5.86, 5.64
5.58, 5.43
5.76, 5.55
5.44, 5.34
ꢁ5.14
ꢁ4.85
ꢁ5.09
ꢁ4.78
0.130
0.228
0.141
0.142
64
19
78
42 (13)
62 (10)
40 (18)
3 (0.3)
15 (0.4)
6.4 (1.3)
19 (2.0)
99 (2.5)
41 (8.3)
a
Vertical ionization potentials calculated at the B3LYP/6-31G** level. The experimental ionization potentials are presented in italics. b Energies of the
HOMO orbitals obtained at the B3LYP/6-31+G* level. c Intramolecular reorganization energy (li) values for 1–4. d Examples of electronic couplings
(absolute values) between the HOMO orbitals in model dimers 1c–4c (in parentheses values for 2b–4b) (Fig. 7) calculated at the uB97X-D/6-31G** level.
e
Hole-transport rate constants (at zero electric field) calculated from the Marcus–Levich–Jortner equation. Constant values of ls ¼ 0.3 eV and DGꢂ ¼
0 are considered for all the compounds. f Hole mobility estimations calculated on the frame of a perfectly symmetric one dimensional model as described
ꢂ
ꢂ
in ref. 12. A fixed distance of 4.5 A between the monomers is considered in the case of 1 and 5.5 A for 2–4. The value of the electric field used in these
calculations is 106 V cmꢁ1. In parentheses are reported the kHT and mHT values corresponding to t values shown in parentheses.
3022 | J. Mater. Chem., 2012, 22, 3015–3027
This journal is ª The Royal Society of Chemistry 2012

View Article Online
amorphous layer of 4 reaches 1.2 ꢀ 10ꢁ3 cm2 Vꢁ1 sꢁ1 at the
electric field of 106 V cmꢁ1 (Table 4). Ortho-Substituted derivative
2 shows only slightly inferior charge-transporting properties,
while the meta-substituted compound shows the lowest hole
mobility among methoxy-substituted derivatives.
As for the trend of the li values for 1–4, multiple factors may
be involved and clear explanation cannot be given. In the
following only three factors will be considered: (i) the factor
discussed above can influence the relative li values for 1–4. The
HOMO–HOMO-1 splitting (Table S1†) is the greatest for 2
(0.230 eV) and the smallest for 4 (0.140 eV). The change in the
TPA/TPA distance upon oxidation also follows the same trend
(ꢁ0.074 A and ꢁ0.030 A for 2 and 4 respectively, Fig. S1†),
suggesting consequently stronger influence of this factor in the
case of compound 2. (ii) The change in the Ph–‘‘N-plane’’ dihe-
dral angles upon oxidation is the largest for 2 (36ꢂ) and the
smallest for 4 (10.5ꢂ). The change in the N–C bond-lengths, for
In order to understand the effect of the methoxy groups on the
hole mobilities different factors could be evocated. In the
following we focus only on some molecular parameters influ-
encing the hole mobility in the amorphous bulk materials. These
aspects will be discussed in terms of hole-transfer rate-constants
(eqn (1)) and two parameters entering it, intramolecular reor-
ganization energy (li) and electronic coupling (t). Estimated
mobilities, based on our kHT values and a simple one dimensional
model,12 will also be presented. Other effects like morphological
differences in the amorphous film, defects, and charge-traps are
out of the scope of this work.
ꢂ
ꢂ
ꢂ
instance, is consequently the most important for 2 (ꢁ0.046 A)
ꢂ
and the smallest for 4 (ꢁ0.015 A), with the greatest contribution
coming in each case from the N–C bond with the internal (non-
substituted) phenyl ring (Fig. S1, ESI†). In the case of 1 and 3 this
ꢂ
ꢂ
parameter exhibits intermediate values (ꢁ0.025 A and ꢁ0.023 A
respectively). (iii) The anti-bonding contributions of the C(Ph)–
O bonds in the HOMO orbitals are reduced upon oxidation.
These contributions are absent in the case of 1, they are negligible
in the case of 3 (negligible contribution from oxygen atoms in the
HOMO, Fig. 4e), but are the strongest ones for 4 (Fig. 4i). The
changes in the C–O bond-lengths upon oxidation seem to
support this idea: the C–O bonds are shortened more impor-
3.6.1 Intramolecular reorganization energies. The li values
for compounds 1–4 are presented in Table 3 and are found to
increase in the order 1 < 3 z 4 < 2. In order to get some insight
on this trend we firstly compared the li values of 1 (0.13 eV) and
TPA (0.11 eV).42 The slightly larger li value for 1 seems to be in
disagreement with the idea that the larger space extension of the
p system in 1 should correspond to reduced reorganization
energy. One possible factor explaining this discrepancy may be
due to the presence of the cyclohexylidene bridge in 1, as sug-
ꢂ
ꢂ
tantly in the case of 4 (ꢁ0.015 A) than for 3 (ꢁ0.011 A). As for
compound 2, smaller changes in the C–O bond lengths are found
ꢂ
6
gested by the greater li value of TPA–CH3 (0.126 eV as
(ꢁ0.003 A) probably due to the important O/Ph hindrance
compared to 0.11 eV for TPA). Replacing one H atom in TPA–
CH3 with a second TPA results in a model compound (TPA–
CH2–TPA) for which, in agreement with the common ideas, we
found a smaller li value (0.098 eV) as compared to TPA.
However, going from TPA–CH2–TPA to 1 (replacing the CH2
bridge by the cyclohexylidene one) results in increased li from
0.098 to 0.13 eV. This should be mainly due to the difference in
the through-space interaction between the two TPA moieties in 1
and TPA–CH2–TPA. Indeed, the TPA/TPA interaction should
be non-negligible as suggested by the shortest TPA/TPA
(Fig. S1†).
All these factors are expected to contribute to the increasing
reorganization energy. The great importance of the first two
factors for 2 gives a possible explanation for the largest li value
for this compound, whereas the smaller and roughly similar
influence of these factors for 1, 3, and 4 could explain the cor-
responding smaller and similar li values.
It is worth noting that the factors discussed above are
considered as acting independently, thus ignoring the possible
couplings between them and also with other factors not consid-
ered here. However, we remember that the aim of the above
qualitative analysis is simply to identify some of the factors
influencing the li values.
ꢂ
ꢂ
distances (2.49 A and 2.56 A for 1 and TPA–CH2–TPA respec-
tively). This assumption is also supported by the non-negligible
splitting between HOMO-1 and HOMO orbitals (0.156 eV and
0.100 eV for 1 and TPA–CH2–TPA respectively), these last
orbitals corresponding to the bonding and anti-bonding combi-
nations of the local HOMO (TPA) orbitals (Fig. S3, ESI†). In the
case of TPA–CH2–TPA model compound, this interaction is
thus weaker as compared to 1 and the change in the TPA/TPA
We now turn to the correlation of the li values with the
experimental results. In view of eqn (1) and considering fixed
identical values for the parameters t and ls, it can be expected
that the kHT values decrease in the order 1 > 3 z 4 > 2, which is
different from the experimental trend 4 > 2 > 3 > 1 of the
mobility values (note that somewhat speculative parallel is made
here between kHT and the mobility values). This result also seems
to be different from previous studies43 having found that in
a family of similar TPA compounds the hole mobilities follow the
ꢂ
distance upon oxidation is also smaller (ꢁ0.035 A for TPA–
ꢂ
CH2–TPA as compared to ꢁ0.044 A for 1), which can explain the
smaller influence (smaller sensitivity upon oxidation) of this
interaction on the li value of TPA–CH2–TPA.48
Table 4 Hole mobility values in amorphous layers of 1–4. Magnitudes in parentheses are of mixtures with PC-Z (mass ratio 1 : 1)
Sample
Layer thickness/mm
m0/cm2 Vꢁ1 sꢁ1
m/cm2 Vꢁ1 sꢁ1 at 106 V cmꢁ1
1
2
3
4
(5.0)
(4.9 ꢀ 10ꢁ9
)
(2.0 ꢀ 10ꢁ6
8.2 ꢀ 10ꢁ4
)
2.0 (5.6)
2.8 (8.0)
4.0 (5.4)
6.5 ꢀ 10ꢁ5 (7.8 ꢀ 10ꢁ6
5.0 ꢀ 10ꢁ6 (2.6 ꢀ 10ꢁ9
3.6 ꢀ 10ꢁ5 (7.7 ꢀ 10ꢁ7
)
)
)
4.0 ꢀ 10ꢁ4 (1.0 ꢀ 10ꢁ5
1.2 ꢀ 10ꢁ3 (2.4 ꢀ 10ꢁ5
)
)
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 3015–3027 | 3023

View Article Online
same trend as the li values. It seems that the definition ‘‘family of
similar compounds’’ is subtle and could not apply to our
compounds 1–4.
the non-negligible interaction energies found in the case of 2a–4a
(6–7 kcal molꢁ1). Similar C–H/p interactions have been previ-
ously calculated theoretically or have been observed in crystal
structures.49–53 (iii) In the absence of the methoxy groups
(compound 1), the interactions between the phenyl groups of two
adjacent monomers should involve C(Ph)–H/p(Ph) short
contacts. The optimized structure for one dimer of this type is
shown in Fig. 7 (model compound 1c) where short H/C
3.6.2 Electronic couplings. In order to shed light on this
disagreement we focus on the electronic coupling properties of 1–
4. In this respect, different dimers were optimized for three model
compounds: (i) dimers of ortho-, meta- and para-methoxy
substituted aniline, referred hereafter as 2a, 3a, and 4a (Fig. 7),
(ii) dimers of the ortho-, meta- and para-methoxy substituted
TPA referred hereafter as 2b, 3b, and 4b (containing one non-
substituted phenyl ring, Fig. 7). In both model compounds the
optimized geometries show short C–H/p(Ph) contacts between
a methoxy group of one monomer and the p(Ph)-system of the
second one. These C–H/p(Ph) interactions seem to be suffi-
ciently strong for retaining the dimers bonded, as suggested by
ꢂ
contacts ranging between ꢃ2.6 and 2.9 A were found. Starting
from this geometry and introducing methoxy groups in ortho-,
meta- and para-positions, geometries equivalent to 1c were also
optimized (referred hereafter as 2c–4c), where C–H/O, C–H/p
(Ph), and C(Ph)–H/p(Ph) short contacts can be observed
(Fig. 7).
The electronic couplings between the monomer-HOMOs for
each dimer were subsequently calculated. In the case of the
Fig. 7 Optimized geometries for some model dimers: (2a–4a) ortho-, meta- and para-methoxy substituted aniline dimers respectively; (2b–4b) ortho-,
meta- and para-methoxy substituted TPA dimers respectively; (1c) TPA model dimer; (3c and 4c) dimers of meta- and para-methoxy substituted TPA
corresponding to equivalent geometries as compared to TPA dimer 1c (see text for details). Some relevant C–H/p(Ph), C–H/O, and C–H/N short
ꢂ
contacts ranging from 2.6–2.8 A are highlighted (dashed lines in black, red, and blue colours respectively). The dN–N corresponds to the distance between
the N atoms.
3024 | J. Mater. Chem., 2012, 22, 3015–3027
This journal is ª The Royal Society of Chemistry 2012

View Article Online
aniline model dimers (2a–4a, Fig. 7) electronic couplings of 0.108
eV, 0.001 eV, and 0.077 eV respectively were found, indicating
for the meta-substituted isomer (3a) a value of ꢃ1 to 2 orders of
magnitude smaller than for the ortho- and para-isomers. The
smaller coupling value for the meta-substituted compound can be
understood by considering the geometries of these model dimers.
Indeed, the short contacts highlighted in Fig. 7 for 3a and 4a
suggest that the only possible overlap between the monomer-
HOMOs in these geometries can be established between the p
(Ph) part of one monomer and the oxygen p atomic orbital of the
other monomer. The contribution in the HOMO orbital from the
oxygen atoms in the meta-position is negligible (Fig. 4e), result-
ing in smaller overlap and electronic coupling, as compared to
the non-negligible contribution from the oxygen atoms in the
case of para- and (to a lesser extent) ortho-isomers (encircled
parts in Fig. 4i and c).
spacing between monomers. While model dimers 2b–4b may
correspond to this type of molecular packing, the geometries of
1c–4c mostly correspond to a compact packing (Fig. 7). Indeed,
the distance between the monomer centers in 1c–4c is found to
ꢂ
ꢂ
range between ꢃ5 and 6 A as compared to ꢃ7 and 8 A for 2b–4b.
It can be then assumed that in the case of ‘‘dilute’’ films for
compound 1, dimer geometries similar to 1c cannot be estab-
lished or should involve large mean distance, which suggests very
small electronic couplings. On the contrary, despite the larger
center-to-center mean distances, dimer geometries similar to 2b–
4b (Fig. 7) should still be established between the methoxy-
substituted compounds, giving consequently a possible expla-
nation for the higher hole mobility values of 2–4 in ‘‘dilute’’ films
as compared to 1.
As for the films of pure compounds, one can reasonably
assume that the packing mode between adjacent molecules in the
amorphous materials cannot be entirely compact. The hole-path
could then be figured out as an alternation of compact (similar to
2c–4c) and dilute (similar to 2b–4b) packed dimers. In this case
we assume that the trend of the mobility values will be dominated
by the dilute packed dimers, which suggests small spread of the
mobility values with the best one expected for compound 4
(Table 3, values in parentheses).
In the case of the TPA-based model compounds 2b–4b, the
electronic couplings shown in Table 3 (values in parentheses)
vary in a much smaller range as compared to 2a–4a, probably
due to the larger space extent of the HOMO orbitals but also to
the additional steric hindrance between monomers. However, the
general trend is the same as in the case of 2a–4a, with the smallest
t value found for 3b (0.01 eV) and the largest one for 4b (0.018
eV). As for model dimers 1c–4c (Fig. 7), much stronger electronic
couplings ranging from 0.04–0.06 eV are found as compared to
2a–4a and 2b–4b. The values are however similar between 1c and
4c, probably due to the principal role of the TPA core in these
dimer geometries. It is interesting to point out that the strongest
couplings are calculated for non-substituted compound 1c,
which seems to agree with previous observations.7 However,
while this order might be reversed in the frame of a thorough
study, the significant result is that the couplings for 1c–4c are of
similar values.
Admittedly, only a small number of model dimers taken at
idealized static geometries were considered and film
morphology and other factors were ignored in this analysis,
meaning that direct comparison between the calculated kHT
(and mHT) and the experimental mobilities cannot be proposed.
However, these results point to the special role of the methoxy
groups at the microscopic level, giving possible explanations
on: (i) why methoxy substituted compounds 2, 3, and 4 show
higher kHT values in mixtures with PC-Z as compared to the
non-substituted one; (ii) why the para-substituted compound
shows the highest kHT value and (iii) why the meta-substituted
compound in ‘‘dilute’’ film shows lower kHT value than the
para-isomer.
It is worth noting that compounds 1–4 in the radical cationic
state should exhibit mixed valence (MV) character. The intra-
molecular electronic couplings between the two TPA moieties,
estimated as half of the energy difference between HOMO and
HOMO-1 orbitals in the neutral species, range between 0.07 and
0.12 eV, thus being stronger than the intermolecular electronic
couplings shown in Table 3. In these conditions we assume that
among the intra- and intermolecular hole-transfer steps the rate-
limiting one is the intermolecular step.54
Conclusions
In this work we have synthesized three new isomeric 1,1-bis(4-
aminophenyl)cyclohexane based triphenylamine containing
methoxy groups in different positions and have studied their
thermal, optical, photophysical and photoelectrical properties.
Methoxy-substituted derivatives of 1,1-bis(4-(N,N-diphenyl)
aminophenyl)cyclohexane show lower ionization potentials and
higher hole drift mobilities than the compound containing no
OCH3 groups. Among methoxy-substituted derivatives the para-
substituted compound shows the lowest ionization potential
(5.34 eV) and the highest hole mobility which reaches 1.2 ꢀ 10ꢁ3
3.6.3 Hole-transfer rate constants. The intermolecular hole-
transfer rate constants (kHT, Table 3) have been calculated in the
frame of Marcus–Levich–Jortner equation55 by using the li
values for 1–4 and the electronic coupling values for 2b–4b and
1c–4c56 (see ESI† for more details). Hole-mobility values esti-
mated on the frame of a perfectly symmetric one dimensional
model12 are also presented in Table 3. As expected, higher
mobility values are calculated as compared to the experimental
ones, which is due to the limited and idealized frame of the model
applied in this study. However, the significant result is that
similar kHT (and mHT) values are found for 2–4, differing in each
case (2b–4b or 2c–4c) by less than one order of magnitude, which
seems to correspond to the experimental results.
cm2 Vꢁ1 sꢁ1 at the electric field of 106 V cmꢁ1
.
The theoretical analysis is based on the influence of two
factors: (i) the comparison of hole-transfer rate constants in the
amorphous bulk material by considering the simultaneous
influence of two electronic parameters, the intramolecular reor-
ganization energy and the electronic coupling. Our results
suggest that the trend of the kHT values is dominated by the
electronic coupling parameter whilst the use of li values sepa-
rately suggests in our case the wrong mobility trend. (ii) The
comparison of the HOMO level energies, which are strongly
It is worth noting that the comparison of the experimental
mobilities between 1 on the one hand and 2–4 on the other hand
concerns only mixtures 1 : 1 with PC-Z, implying ‘‘dilute’’
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 3015–3027 | 3025

View Article Online
12 Y. Olivier, V. Lemaur, J. L. Bredas and J. Cornil, J. Phys. Chem. A,
2006, 110, 6356.
related to the efficiency of the film–electrode contacts in the
DSSCs devices.
13 R. A. Marcus, J. Chem. Phys., 1956, 24, 966.
14 V. G. Levich, Adv. Electrochem. Sci. Eng., 1966, 4, 249.
15 R. A. Marcus and N. Sutin, Biochim. Biophys. Acta, 1985, 811, 265.
16 R. A. Marcus, Rev. Mod. Phys., 1993, 65, 599.
17 W. Kohn and L. Sham, Phys. Rev., 1965, 140, A1133.
18 M. Kirkus, R. Lygaitis, M. H. Tsai, J. V. Grazulevicius and C. C. Wu,
Synth. Met., 2008, 158, 226.
Both factors (electronic couplings and HOMO energies) work
in concert, suggesting enhanced performances of the methoxy-
substituted compounds with the highest one found for the
methoxy-para-substituted compound.
In the frame of this study, the role of the methoxy groups is
found to be related to two principal properties: the well-known
mesomeric (p-donor) effect and the possibility to establish
C–H/p(Ph) and C–H/X (X ¼ O, N) hydrogen bonds. The
effects of these properties are multiple: (i) destabilize the HOMO
orbital and decrease the Ip values of 2, 3, and 4 with respect to 1.
(ii) Introduce anti-bonding interactions in the HOMO of
para- and (to a lesser extent) ortho-substituted compounds which
are reduced in the excited states. Due to this effect, the para- (and
ortho-) substituted compounds undergo supplementary geom-
etry-relaxation effects after photon absorption, suggesting one
possible contribution to the increased Stokes shifts. The same
effect is found to contribute to the increase of the intramolecular
reorganization energies of para- (and ortho-) substituted
compounds. (iii) In these amorphous materials, where compact
and dilute packed contacts between monomers coexist, the
methoxy groups introduce additional possibilities for establish-
ing short contacts between adjacent molecules (at larger center-
to-center distances as compared to the non-substituted
compound) which seem to play a central role in the case of dilute
packing. In this frame, the electronic couplings and kHT for the
methoxy-substituted compounds are enhanced as compared to
the non-substituted one.
19 E.
Miyamoto,
Electrophotography, 1989, 28, 364.
Y.
Yamaguchi
and
M.
Yokoyama,
20 E. Montrimas, V. Gaidelis and A. Pazera, Lith. J. Phys., 1966, 6, 569.
21 S. M. Vaezi-Nejad, Int. J. Electron., 1987, 62, 361.
22 S. Grigalevicius, V. Getautis, J. V. Grazulevicius, V. Gaidelis,
V. Jankauskas and E. Montrimas, Mater. Chem. Phys., 2001, 72, 395.
23 C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785.
24 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
25 M. J. Frisch, G. W. Trucks and H. B. Schlegel, GAUSSIAN 09,
Revision B.01, Gaussian, Inc., Wallingford, CT, 2009.
26 E. K. U. Gross and W. Kohn, Phys. Rev. Lett., 1985, 55, 2850.
27 E. Runge and E. K. U. Gross, Phys. Rev. Lett., 1984, 52, 997.
28 E. K. U. Gross and W. Kohn, Adv. Quantum Chem., 1990, 21, 255.
29 R. Bauernschmitt and R. Ahlrichs, Chem. Phys. Lett., 1996, 256, 454.
30 M. E. Casida, C. Jamorski, K. C. Casida and D. R. Salahub, J. Chem.
Phys., 1998, 108, 4439.
31 J. L. Bredas, D. Beljonne, V. Coropceanu and J. Cornil, Chem. Rev.,
2004, 104, 4971.
32 M. D. Newton, Chem. Rev., 1991, 91, 767.
33 K. Senthikumar, F. C. Grozema, F. M. Bickelhaupt and
L. D. A. Siebbeles, J. Chem. Phys., 2003, 119, 9809.
34 E. F. Valeev, V. Coropceanu, D. A. da Silva, S. Salman and
J. L. Bredas, J. Am. Chem. Soc., 2006, 128, 9882.
35 H. Li, J. L. Bredas and C. Lennartz, J. Chem. Phys., 2007, 126,
164704.
36 T. Kawatsu, V. Coropceanu, D. A. da Silva, S. Salman and
J. L. Bredas, J. Phys. Chem. C, 2008, 112, 3429.
37 J. D. Chai and M. Head-Gordon, Phys. Chem. Chem. Phys., 2008, 10,
6615.
38 G. Sini, J. S. Sears and J. L. Bredas, J. Chem. Theory Comput., 2011,
7, 602.
39 K. Yesudas, G. Sini and J. L. Bredas, Manuscript under preparation.
40 S. F. Boys and F. Bernardi, Mol. Phys., 1970, 19, 553.
41 F. Ullmann and J. Bielecki, Ber. Dtsch. Chem. Ges., 1901, 34, 2174.
42 M. Malagoli and J. L. Bredas, Chem. Phys. Lett., 2000, 327, 13.
43 K. Sakanoue, M. Motoda, M. Sugimoto and S. Sakaki, J. Phys.
Chem. A, 1999, 103, 5551.
Finally, the absorption bands of these compounds are situated
in the UV domain and the thermal properties are satisfying
which makes them good candidates for DSSC applications.
Acknowledgements
This research was funded by a grant no. MIP-059/2011 from the
Research Council of Lithuania. G.S. thanks Dr Veaceslav
Coropceanu and Prof. Dr Bernard Kippelen (Georgia Institute
of Technology) for stimulating discussions.
44 B. C. Lin, C. P. Cheng and Z. P. M. Lao, J. Phys. Chem. A, 2003, 107,
5241.
45 H. Xiao, B. Leng and H. Tian, Polymer, 2005, 46, 5707.
46 The effect of the para-methoxy substitutions on the Ph–‘‘N–plane’’
dihedral angles and on the Ph–N–Ph p-conjugation should be
cancelled out due to the opposite signs of these changes (see Section
3.3).
References
47 It is worth remembering that this explanation ignores the orbital level
shift happening at the material–electrode interface. However, due to
the structural similarity of these compounds no profound alteration
of the above trend should be expected.
48 The contribution of the bridge in the relaxation energy should be
negligible as suggested from the comparison between the li values
of TPA–CH3 and TPA–C6H11 model compound (0.126 and 0.13 eV
respectively, this work).
49 T. Takagi, A. Tanaka, S. Matsuo, H. Maezaki, M. Tani, H. Fujiwara
and Y. Sasaki, J. Chem. Soc., Perkin Trans. 2, 1987, 1015.
50 J. L. Bredas and G. B. Street, J. Chem. Phys., 1989, 90,
7291.
51 D. Braga, F. Grepioni and E. Tedesco, Organometallics, 1998, 17,
2669.
1 Y. J. Shirota, J. Mater. Chem., 2005, 15, 75.
2 S. Gunes, H. Neugebauer and N. S. Sariciftci, Chem. Rev., 2007, 107,
1324.
3 J. V. Grazulevicius, Polym. Adv. Technol., 2006, 17, 694.
4 P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for
Xerography, Marcel Dekker, New York, 1998.
5 E. T. Seo, R. F. Nelson, J. M. Fritsch, L. S. Marcoux, D. W. Leedy
and R. N. Adams, J. Am. Chem. Soc., 1966, 88, 3498.
6 J. H. Pan, Y. M. Chou, H. L. Chiu and B. C. Wang, Aust. J. Chem.,
2009, 62, 483.
7 J. L. Maldonado, M. Bishop, C. Fuentes-Hernandez, P. Caron,
B. Domercq, Y. D. Zhang, S. Barlow, S. Thayumanavan,
M. Malagoli, J. L. Bredas, S. R. Marder and B. Kippelen, Chem.
Mater., 2003, 15, 994.
52 L. Brunel, F. Carre, S. G. Dutremez, C. Guerin, F. Dahan,
O. Eisenstein and G. Sini, Organometallics, 2001, 20, 47.
53 L. Orian, P. Ganis, S. Santi and A. Ceccon, J. Organomet. Chem.,
2005, 690, 482.
54 The geometries of 1–4 are not symmetric and part of the energy
difference between HOMO and HOMO-1 orbitals might be due to
the mutual polarization between the two TPA moieties. However,
the two geminal Ph rings in 1–4 are positioned almost
8 R. D. Hreha, C. P. George, A. Haldi, B. Domercq, M. Malagoli,
S. Barlow, J. L. Bredas, B. Kippelen and S. R. Marder, Adv. Funct.
Mater., 2003, 13, 967.
9 X. Wu, A. P. Davis, P. C. Lambert, L. K. Steffen, O. Toy and
A. J. Fry, Tetrahedron, 2009, 65, 2408.
10 H. Bassler, Phys. Status Solidi B, 1993, 175, 15.
11 V. Coropceanu, J. Cornil, D. A. da Silva Filho, Y. Olivier, R. Silbey
and J. L. Bredas, Chem. Rev., 2007, 107, 2165.
3026 | J. Mater. Chem., 2012, 22, 3015–3027
This journal is ª The Royal Society of Chemistry 2012

View Article Online
face-to-facebetween them and the role of the mutual polarization in
the HOMO–HOMO-1 energy splitting should be small. Test
calculations, where the two geminal groups are forced to be
positioned as symmetrically between them as possible, result in
practically the same HOMO–HOMO-1 energy splitting.
Accordingly, we conclude that the intramolecular couplings in 1–4
are still more important than the intermolecular ones (see also the
discussion in Section 3.6.3). For more details on the couplings in
the M.V. compounds, see: V. Coropceanu, M. Malagoli,
J. M. Andre and J. L. Bredas, J. Am. Chem. Soc., 2002, 124, 10519.
55 J. Jortner, J. Chem. Phys., 1976, 64, 4860.
56 Test calculations using the li values of 2b–4b (0.304 eV, 0.170 eV, and
0.271 eV respectively) do not modify the following discussion and
conclusions. It is also assumed that adding the cyclohexylidene unit
and the second TPA to model dimers 2b–4b and 1c–4c should not
change the trends on the t values.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 3015–3027 | 3027