Structure engineering of naphthalene diimides for improved charge carrier mobility: Self-assembly by hydrogen bonding, good or bad?

Article abstract of DOI:10.1039/c2jm32554c

Two families of naphthalene diimide (NDI) derivatives were compared and contrasted for the effect of self-assembly on charge carrier transport. One series of NDI derivatives had a terminal phenyl ring attached to a hexyl spacer substituted naphthalene core either through an ester or an amide linkage (NDI-E and NDI-A, respectively), while the other series had a 3,4,5-tridodecyloxy phenyl unit (NDI-E3, NDI-A3) instead of the terminal phenyl unit. Solution processed thin films of these molecules exhibited n-type charge transport characteristics in a bottom gate top contact organic field effect transistor (OFET) geometry. The amide derivatives showed evidence of self-organization with observation of red shifted aggregate emission in solution as well as solid state. Variable temperature FTIR studies in the solid state confirmed the existence of strong hydrogen bonding which could be broken only at very high temperature. However, contrary to expectations, the NDI ester derivatives showed better device efficiency with electron mobilities in the range 8.5 × 10^-3 to 2 × 10^-2 cm² V^-1 s^-1 and on/off ratio ～10⁴. The thin film crystallinity and morphology of NDI-E and NDI-A were examined through X-ray diffraction and atomic force microscopy (AFM). The correlation of crystallinity, hydrogen bonding and charge carrier mobility was studied using energy minimized structures from density functional theory (DFT). The higher electron mobility of ester linked NDI derivatives over the amide linked ones was attributed to the freedom in charge transport pathways offered by a three dimensional crystalline organization in the ester compared to the restricted directional hydrogen bonding interaction in the amide derivatives. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2jm32554c

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 15235
www.rsc.org/materials
PAPER
Structure engineering of naphthalene diimides for improved charge carrier
mobility: self-assembly by hydrogen bonding, good or bad?†
a
a
b
a
Nagesh B. Kolhe, R. Nandini Devi, Satyaprasad P. Senanayak, B. Jancy, K. S. Narayan and S. K. Asha
b
a
*
Received 23rd April 2012, Accepted 30th May 2012
DOI: 10.1039/c2jm32554c
Two families of naphthalene diimide (NDI) derivatives were compared and contrasted for the effect of
self-assembly on charge carrier transport. One series of NDI derivatives had a terminal phenyl ring
attached to a hexyl spacer substituted naphthalene core either through an ester or an amide linkage
(NDI-E and NDI-A, respectively), while the other series had a 3,4,5-tridodecyloxy phenyl unit (NDI-E3,
NDI-A3) instead of the terminal phenyl unit. Solution processed thin films of these molecules exhibited
n-type charge transport characteristics in a bottom gate top contact organic field effect transistor
(OFET) geometry. The amide derivatives showed evidence of self-organization with observation of red
shifted aggregate emission in solution as well as solid state. Variable temperature FTIR studies in the
solid state confirmed the existence of strong hydrogen bonding which could be broken only at very high
temperature. However, contrary to expectations, the NDI ester derivatives showed better device
efficiency with electron mobilities in the range 8.5 ꢀ 10ꢁ3 to 2 ꢀ 10ꢁ2 cm2 Vꢁ1 sꢁ1 and on/off ratio ꢂ104.
The thin film crystallinity and morphology of NDI-E and NDI-A were examined through X-ray
diffraction and atomic force microscopy (AFM). The correlation of crystallinity, hydrogen bonding
and charge carrier mobility was studied using energy minimized structures from density functional
theory (DFT). The higher electron mobility of ester linked NDI derivatives over the amide linked ones
was attributed to the freedom in charge transport pathways offered by a three dimensional crystalline
organization in the ester compared to the restricted directional hydrogen bonding interaction in the
amide derivatives.
in the form of different degrees of crystallinity, morphology at
the dielectric interface, grain boundaries, twining and grain
connectivity.12 All these strongly affect the electrical transport
property of the resultant film as multiple length scales are
involved in the electrical transport in such systems. The charge
motion in these films has a hopping like transport at the dielec-
tric–semiconductor interface, which has a higher density of
defect states, and band like transport is dominant in the ordered
bulk of the crystalline thin films.13,14 Thus, bottom-up
approaches designed for enabling microscopic control over
crystal engineering have been the focus of increasing attention in
this area of research, where the success is often gauged by the
enhancement of the microscopic mobility. However crystal
engineering is still a matter of trial and error and a simple
structural variation has been shown to have a huge impact on the
way a molecule organizes in the solid state. A face-to-face
arrangement of molecules may enhance the p–p overlap,
however charge transport seems to be further enhanced with
slipped crystal packs since 2-D structure is preserved.15 The
rational design of molecules with improved optoelectronic
properties should take into account their geometrical as well as
electrical structures. For instance, a pentacene crystal has
a herringbone structure with an angle of 52ꢃ between molecular
Introduction
Organic n-type semiconductors with high charge-carrier mobil-
ities have been increasingly reported in the literature in recent
years.1–9 A fundamental requirement in the development of new
materials with improved properties is the challenging task of
achieving a rigorous understanding of the relationship between
molecular structure and charge transport properties. Efficient
charge transport has been associated with better packing leading
to strong electronic coupling between neighboring molecules.10
However, attaining high charge carrier mobilities is not just
a question of simple design principles. It is known that for the
same organic molecule, the observed charge carrier mobilities
vary depending on the mode of processing – vacuum deposited,
spin coated or single crystal.11 Defects are observed even using
controlled growth deposition techniques of crystalline thin-films
aPolymer Science & Engineering Division, National Chemical Laboratory,
Pune-411008, Maharashtra, India. E-mail: sk.asha@ncl.res.in; Fax: +91-
20-25902615
bJawaharlal Nehru Centre for Advanced Scientific Research, Jakur,
Bangalore 560064, Karnataka, India
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c2jm32554c
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 15235–15246 | 15235

View Online
planes whereas perfluoropentacene is packed nearly perpendic-
ular along the same direction.16 Perfluoropentacene has exhibited
electron transport with mobilities ꢂ0.22 cm2 Vꢁ1 sꢁ1 in FET
devices whereas pentacene exhibits FET hole mobilities in the
range of 0.5–1 cm2 Vꢁ1 sꢁ1. In these examples, both band gap
engineering of the HOMO–LUMO gap and crystal engineering
of face-to-face p-stacking play a role in the nature as well as
efficiency and stability of charge transport. Among the n-type
materials other than C60 derivatives both naphthalene and per-
ylene diimides (NDI–PDI) have been studied quite exten-
sively.17–28 In small molecules based on naphthalene diimide
(NDI) or perylene diimide (PDI) also, the effect of small struc-
tural variations on the molecular packing and thin film
morphology has been intensely looked into. For instance Shukla
et al. described the effect of a simple structural variation like that
of a cyclohexyl end group on the overall crystalline packing and
charge transport characteristics of these NDI derivatives were
compared in an organic field effect transistor (OFET) in bottom
gate top contact geometry. A better understanding of the struc-
ture–device performance relationship in the NDI/PDI class of
compounds is essential to be able to design better molecules as
well as to extract the best device performance from them.
Experimental section
Materials
1,4,5,8-Naphthalenetetracarboxylicdianhydride (NTCDA), 1,6-
diaminohexane, 6-aminohexanol, N,N dimethylacetamide
(DMAc), 3,4,5-trihydroxy benzoic acid and benzoyl chloride
were purchased from Sigma Aldrich and used without further
purification. Zinc acetate, imidazole, thionyl chloride and pyri-
dine were purchased locally and purified using standard
procedures.
charge carrier mobility.29 They demonstrated
a dramatic
improvement in FET mobility of ꢂ6 cm2 Vꢁ1 sꢁ1 which was
attributed to the propensity of the cyclohexyl terminated NDI
derivatives to assemble into thin films with bulk phase crystalline
packing. The possibility of using non-covalent interactions like
p-stacking and H bonding to drive the molecule to pack in the
favored arrangement for improved charge carrier transport is an
interesting question to address. A few reports are available in the
literature where researchers have sought answers to the above
question.30–36 An example of crystal engineering taking help from
Measurements
1H NMR spectra were recorded using a 200 MHz Brucker NMR
spectrophotometer with samples dissolved in CDCl3 containing
small amounts of TMS as internal standard. The purity of the
compounds was determined by elemental analysis, as well as
MALDI-TOF. Elemental analysis was performed using a Ther-
mofinnigan flash EA 1112 series, CHNS analyser. The matrix
used for MALDI was 2,5-dihydroxy benzoic acid dissolved in
CHCl3 which was spotted on the MALDI target and allowed to
dry before introduction into the mass spectrometer. The purity of
the oligomers was further analyzed by SEC (size exclusion
chromatography) in chloroform using polystyrene standards for
the calibration. The flow rate of CHCl3 was maintained as 1 ml
minꢁ1 throughout the experiments, and the sample solutions at
concentrations of 2–3 mg mlꢁ1 were filtered and injected for
recording the chromatograms at 30 ꢃC. FTIR spectra were
recorded using a Brucker a-T spectrophotometer in the range of
4000 to 400 cmꢁ1. UV-Vis spectra were recorded using a Perkin
Elmer Lambda 35 UV-Vis spectrometer. The emission studies
were performed by SPEX Fluorolog F112X spectrofluorimeter.
The fluorescence quantum yields of the naphthalene derivatives
were determined in chloroform using quinine sulfate in 0.5 M
sulfuric acid (f ¼ 0.546) as the standard by exciting at 360 nm.
The optical density at l382 was maintained at 0.1 ꢄ 0.05 to avoid
re-absorption artifacts. For recording solid state photophysical
properties, thin films were spin coated (1000 rpm per 60 s) from
chloroform as well as o-DCB (ortho-dichlorobenzene) solution
on a quartz plate. Thermogravimetric analysis (TGA) was per-
formed using a PerkinElmer STA 6000 thermogravimetric ana-
lyser. Samples were run from 40 to 800 ꢃC with a heating rate of
10 ꢃC minꢁ1 under nitrogen. DSC (differential scanning color-
imeter) measurements were performed on a TA Q10 differential
scanning calorimeter at a heating rate of 10 ꢃC minꢁ1 under
nitrogen atmosphere. Typically, 2–3 mg of sample was placed in
an aluminum pan, sealed properly and scanned from ꢁ50 to
300 ꢃC. The phase behaviors of the molecules were analyzed
using a Leica DM2500P polarized optical microscope equipped
with a Linkam TMS 94 heating and cooling stage connected to a
Linkam TMS 600 temperature programmer. The X-ray diffrac-
€
hydrogen bonding interactions was provided by Wurthner et al.,
where they introduced eight chlorine substituents on the bay
position of a perylene derivative having free NH imide func-
tionalities that could be involved in H bonding.33 The combi-
nation of NH imide group induced H bonding and the
contortion of the ring afforded by the bay substitution resulted in
a two-dimensional p–p-stacked percolation path for electron
transport. The molecule exhibited excellent air-stable operation
with a mobility as high as 0.82 cm2 Vꢁ1 sꢁ1 with Ion/Ioff of 108. In
this example, the supramolecular ordering aided crystallization
in the preferred direction. However, supramolecular organiza-
tion can alter or compete with free crystallization to restrict the
charge transport in all directions. Thus, the question of supra-
molecular ordering versus crystallinity,which is better at inducing
higher charge carrier mobility, is of fundamental importance.
The aim of this work was to address the question of the effect
of three-dimensional crystalline organization vs. directional self-
assembly on electron transport mobility in naphthalene diimide
(NDI) derivatives. Two series of molecules based on NDI were
chosen for this study having identical structure except for
a replacement of ester group with amide. The NDI derivatives
with an ester linkage did not have the additional non-covalent
interaction in the form of hydrogen bonding to direct the
molecular organization, whereas the amide molecules had
a directional hydrogen bonding of the amide linkage as
a constraint to restrict their freedom of organization. The effect
of this minor structural alteration on the self-organizing prop-
erties in solution was probed using UV-Vis absorption and
fluorescence emission studies. The solid state organization was
studied using wide angle X-ray diffraction (WXRD), differential
scanning calorimetry (DSC) and polarized light microscopy
(PLM). Hydrogen bonding in the amide molecules was traced
using variable temperature FTIR studies in the solid state. The
tion data were recorded by using
a Philips analytical
15236 | J. Mater. Chem., 2012, 22, 15235–15246
This journal is ª The Royal Society of Chemistry 2012

View Online
diffractometer using Cu Ka emission, and the spectra were
recorded in the range of (2q) 3–50ꢃ and analyzed using X’pert
software. Powder X-ray diffraction of all the samples was carried
out in a PANalytical X’pert Pro dual goniometer diffractometer.
An X’celerator solid-state detector was employed in wide-angle
excess solvent was removed by distillation under reduced pres-
sure and the slurry was poured into acetone. The obtained
precipitate was filtered and dried in a vacuum oven for 12 hours
(0.1 mm of Hg). The crude compound was purified by column
chromatography using chloroform–methanol (98 : 2 v/v) Yield:
590 mg (82%); 1H NMR (200 MHz, CDCl3) d ppm: 8.73 (s, 4H,
Ar naphthalene), 8.02 (d, 4H, Ar benzoyl), 7.54 (m, 2H, Ar
benzoyl), 7.40, (m, 4H, Ar benzoyl), 4.34 (t, 4H, ester CH2), 4.22
(t, 4H, imide CH2), 1.80–1.54 (16H, CH2). 13C NMR (75 MHz,
CDCl3) d ppm: 166.5, 162.7, 132.7, 130.8, 130.3, 129.4, 128.2,
126.5, 64.8, 40.7, 28.5, 27.8, 26.6, 25.7. FTIR (KBr, cmꢁ1): 2939,
2851,1725, 1660, 1586, 1453, 1371, 1340, 1272, 1245, 1106, 772,
707; Anal. Calcd for C40H38N2O8: C, 71.20; H, 5.68; N, 4.15.
ꢀ
experiments. The radiation used was Cu Ka (1.54 A) with a Ni
filter and the data collection was carried out using a flat holder in
Bragg–Brentano geometry. Care was taken to avoid sample
displacement effects. Variable temperature in situ XRD experi-
ments were carried out in an Anton-Paar XRK900 reactor. AFM
images were taken by a Multimodeꢀ Scanning Probe Micro-
scope equipped with a Nanoscope IVꢀ controller from Veeco
Instrument Inc., Santa Barbara in the tapping mode using a SiN
probe, with maximum scan size 10 mm ꢀ 10 mm and with vertical
range of 2.5 mm. For the AFM studies, samples were prepared by
spin coating from a 1 ꢀ 10ꢁ3 M solution in o-DCB on a glass
slide and allowed to dry before being subjected to AFM analysis.
An FEI-Tecnaiꢀ-F30 electron microscope operating at 100 kV
was used for HR-TEM sample observation, and was equipped
with a Gatan digital camera for recording micrographs. The
electrochemical behavior of NDI ester and amide derivatives was
studied by using a BAS-Epsilon potentiostat.
Found C, 70.76; H, 5.23; N, 4.15%. MALDI-TOF MS (Mw
674); m/z ¼ 697.33 [M + Na]+, 713.28 [M + K]+.
¼
N,N0-Bis[6-hexyl-3,4,5-tridodecyloxybenzester]naphthalene-1,4,5,8-
tetracaroxyldiimide (NDI-E3). The same procedure as above was
followed using (0.5 g, 1.07 mmol) of N,N0-bis[6-hydroxyhexyl]
naphthalene-1,4,5,8-tetracarboxyldiimide, (a) and (1.85 g, 2.67
mmol) 3,4,5-tridodecyloxy benzoyl chloride. The crude
compound was purified by column chromatography using
1
chloroform–methanol (99 : 1 v/v). Yield: 1.3 g (68%). H NMR
(300 MHz, CDCl3) d ppm: 8.73 (S, 4H, Ar naphthalene), 7.22 (s,
4H, Ar benzoyl), 4.28 (t, 4H, ester CH2), 4.20 (t, 4H, imide CH2),
3.99 (t, 12H, OCH2), 1.8–1.24 (m, 136H, CH2), 0.86 (t, 18H,
CH3). FTIR (KBr, cmꢁ1): 2920, 2850, 1717, 1699, 1667, 1581,
1500, 1468, 1430, 1344, 1348, 1245, 1218, 1122, 993 and 770.
Anal. Calcd for C112H182N2O14: C, 75.55; H, 10.30; N, 1.57.
Device fabrication
Bottom-gate top contact structures were fabricated for field
effect transistor studies as per standard procedures.41 In the
present case hydroxyl free divinyltetramethylsiloxanebis-benzo-
cyclobutene (BCB) polymer dielectric was spin coated at 1000
rpm for 1 min and annealed at 290 ꢃC under glove box (N2)
atmosphere. The dielectric thin film thus formed had a capaci-
tance per unit area (C0) ꢂ2 to 4 nF cmꢁ2 measured using
a Keithley 4200 semiconductor parameter analyzer for a film of
thickness 0.5–0.6 mm. The surface of the dielectric was coated
(1500 rpm for 30 s) with a thin monolayer of hexamethyldisila-
zane (HMDS) and annealed at 110 ꢃC for 2 h. This was followed
by spin coating of the solution prepared from 1,2-dichloroben-
zene solution with a concentration of 25 mg mlꢁ1 at 300–400 rpm
for 1 minute to obtain films with thickness ꢂ100 nm. The
spherulite phases in NDI-E3 and NDI-A3 were obtained by
thermal treatment at temperatures above the melting transition
temperature (NDI-A3 ꢂ 150 ꢃC and NDI-E3 ꢂ 100 ꢃC) for 20–30
min and gradual cooling. In the case of the molecules NDI-A and
NDI-E, where the melting transition temperature was very high,
annealing of samples was done at 200 ꢃC and 150 ꢃC respectively
to get the optimum device performance. The average and the
highest mobility magnitude of each system are mentioned in the
table and are based on measurements on a number of devices
(ꢂ10 to15 devices of each molecule).
Found C, 75.51; H, 10.85; N, 1.67%. MALDI-TOF MS (Mw
1780.65); m/z ¼ 1888.82 [M + Ag]+.
¼
N,N0-Bis[6-hexylbenzamide]naphthalene-1,4,5,8-tetracarbox-
yldiimide (NDI-A). (0.5 g, 1.86 mmol) NTCDA, (0.9 g, 4.10
mmol) N-(6-aminohexyl)benzamide (b), zinc acetate (50 mg) and
(20 g) of imidazole were taken ꢃin a 100 ml two neck round
bottom flask and heated to 130 C under nitrogꢃen atmosphere
for 18 h. The reaction mixture was cooled to 90 C and treated
with 150 ml of 2 N hydrochloric acid. The crude product was
filtered, washed with plenty of distilled water followed by
washing with hot methanol and then dried in a vacuum oven for
12 hours (0.1 mm of Hg). The crude compound was purified by
column chromatography using chloroform–methanol (98 : 2
1
v/v). Yield: 890 mg (71%); H NMR (200 MHz, CDCl3) d ppm:
8.78 (s, 4H, Ar naphthalene), 7.71–7.44 (m, 10H, Ar benzoyl),
4.22 (t, 4H, CH2), 3.50 (t, 4H, amide CH2), 1.80–1.54 (16H,
CH2). 13C NMR (75 MHz, CDCl3) d ppm: 163.2, 159.9, 130.8,
130.2, 128.8, 128.1, 69.7, 47.4, 40.6, 31.3, 29.6, 29.5, 29.2. FTIR
(KBr, cmꢁ1): 3299 (NH), 2938, 2852, 1697 (amide CO), 1652,
1586, 1453, 1371, 1340, 1272, 1245, 1106, 772, 707; Anal. Calcd
for C40H40N4O6: C, 71.41; H, 5.99; N, 8.33. Found C, 70.97; H,
5.64; N, 8.77%. MALDI-TOF MS (Mw ¼ 672.29); m/z ¼ 695.36
[M + Na]+, 711.32 [M + K]+.
Synthesis of ester and amide functionalized naphthalenebisimides
N,N0-Bis[6-hexylbenzester]naphthalene-1,4,5,8-tetracarboxyldiimide
(NDI-E). (0.5 g, 1.07 mmol) of N,N0-bis[6-hydroxyhexyl]naph-
thalene-1,4,5,8-tetracarboxyldiimide, (a) in 10 ml dry pyridine
was heated to 80 ꢃC under nitrogen atmosphere to obtain
a homogeneous solution. To this solution benzoyl chloride
(0.33 g, 2.36 mmol) was added dropwise over a period of 15
minutes. The reaction was continued for 12 hours and then the
N,N0-Bis[6-hexyl-3,4,5-tridodecyloxybenzester]naphthalene-
1,4,5,8-tetracaroxyldiimide (NDI-A3). The same procedure as
above was followed using (80 mg, 0.268 mmol) of NTCDA and
(455 mg, 0.589 mmol) N-(6-aminohexyl)-3,4,5-tris(dodecyloxy)
benzamide (b). The crude material was purified by column
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 15235–15246 | 15237

View Online
chromatography using chloroform–methanol (99 : 1 v/v). Yield:
380 g (77%). 1H NMR (300 MHz, CDCl3) d ppm: 8.67 (S, 4H, Ar
naphthalene), 6.89 (s, 4H, Ar benzoyl), 6.07 (broad NH amide),
4.14 (t, 4H, imide CH2), 3.93 (t, 12H, OCH2), 4.38 (t, 4H, amide
CH2), 1.8–1.24 (m, 136H, CH2), 0.86 (t, 18H, CH3). FTIR (KBr,
cmꢁ1): 3272 (amide NH) 2920, 2850, 1701 (amide CO), 1699,
1667, 1581, 1500, 1468, 1430, 1344, 1348, 1245, 1218, 1122, 993
and 770. Anal. Calcd for C112H184N4O12: C, 75.63; H, 10.43; N,
3.15. Found C, 75.82; H, 10.89; N, 3.14%. MALDI-TOF MS
(Mw ¼ 1778.39); m/z ¼ 1886.27 [M + Ag]+.
Results and discussion
Synthesis and characterization
The syntheses of the ester and amide functionalized naphthalene
diimides (NDI) are outlined in Scheme 1 and their structures are
shown in Scheme 2. Four derivatives of naphthalene diimides
were synthesized by varying the type of linkage (ester or amide)
and the flexible alkyl chain attached to the phenyl ring, two
without any terminal alkyl chain and two with 3,4,5-tridodecy-
loxy substitution. The NDI ester derivatives were named NDI-E
(phenyl substitution) and NDI-E3 (1,2,3-tridodecyloxy phenyl
substitution). Similarly the NDI amide derivatives were named
NDI-A and NDI-A3 respectively for the phenyl and 3,4,5-trido-
decyloxy phenyl substitution. A core to periphery approach was
adopted for the synthesis of the ester molecule in which the first
step was the synthesis of a hydroxyl functionalized NDI deriv-
ative by a reaction between 1,4,5,8-naphthalene tetracarboxy-
Scheme 2 Structures of all NDI-ester and NDI-amide molecules.
amide molecules in which the required acid chloride was first
monocoupled with 1,6-diaminohexane to obtain the amine
terminated derivative; two equivalents of which were coupled
with NTCDA in imidazole and zinc acetate to obtain the final
amide functionalized molecules. All molecules were purified by
using column chromatography with chloroform and methanol
as mobile phase. The compounds were fully characterized by
1H, 13C NMR spectroscopy, size exclusion chromatography,
MALDI-TOF mass spectroscopy and elemental analysis. All
NDI derivatives were soluble in common organic solvents like
tetrahydrofuran (THF), chloroform (CHCl3) and toluene. The
labeled 1H NMR spectra of all NDI molecules are given in
Fig. S1†. The SEC analysis showed a single peak for all the
molecules confirming the purity. As expected, the elution time of
the molecules (Fig. S2†) increased with decrease in molecular
weights. The amide and the ester molecule of each series differed
in mass by 2 units, which was well reflected in the SEC plots.
Elemental analysis of the compounds was carried out to confirm
the purity and the observed CHN values were matching with the
calculated values. The samples were subsequently subjected to
MALDI-TOF analysis recorded using 2,5-dihydroxybenzoic
acid and titanium dioxide as the matrix and molecular ion peaks
were obtained for cationic species such as [M + Na+], [M + K+],
and [M + Ag+]. The MALDI spectra of the ester and amide series
are given in Fig. S3†. Thus, the NMR, SEC, MALDI-TOF
spectra and elemental analysis of the molecules confirmed the
structure and high purity of the molecules.
licdianhydride
(NTCDA)
and
two
equivalents
of
6-aminohexanol in DMAc at 160 ꢃC using Zn(OAc)2 as catalyst.
It was further coupled with the corresponding acid chloride in
dry pyridine to obtain the ester functionalized molecules. A
periphery to core approach was adopted for the synthesis of the
Optical and electrochemical properties
The photophysical properties of the naphthalene diimide deriv-
atives were investigated by UV-Vis absorption and fluorescence
emission spectroscopy and electrochemical properties were
studied using cyclic voltammetry (CV). The variation in the
optical properties of the naphthalene diimide ester vs. amide
derivatives was used as a tool to study their different molecular
organization in chloroform. Chloroform being a good solvent
for rigid p-systems like naphthalene diimides, all four NDI
molecules were molecularly dissolved in it. The absorption
spectra of all naphthalene diimides in chloroform (Fig. 1a)
showed three characteristic peaks at 380 nm, 360 nm and 342 nm
Scheme 1 Synthetic route for naphthalene diimide ester and amide
derivatives.
15238 | J. Mater. Chem., 2012, 22, 15235–15246
This journal is ª The Royal Society of Chemistry 2012

View Online
yields were calculated using quinine sulfate as the standard and
the values are given in Table 1. Except for NDI-E, the intensity of
emission was very much quenched for the other three molecules,
which was also reflected in the quantum yield values. Quenching
of fluorescence was observed due to the presence of the 3,4,5-
trialkoxy substitution in NDI-E3 and NDI-A3.39,40 The amide
molecules NDI-A and NDI-A3 (inset) additionally exhibited
a broad aggregated red shifted emission peak at ꢂ550 nm, which
reflected their strong tendency to form aggregates in chloroform.
This red shifted aggregate emission at 550 nm was clearly absent
in the two ester molecules thereby differentiating them from the
amide series of molecules in terms of their self-organization
behavior. The device fabrication (discussed later) of the mole-
cules was carried out using ortho-dichlorobenzene (o-DCB) as
solvent. Therefore, the self-organization characteristics of the
ester and amide molecules were studied in o-DCB also (given in
Fig. S4†). A red-shifted aggregate emission around ꢂ550 nm was
observed for NDI-A in o-DCB also. The emission of the trialkoxy
substituted derivatives NDI-A3 and NDI-E3 was quenched
similar to the observation in CHCl3.
The influence on self-organization induced by hydrogen
bonding of the amides was more apparent in the solid state
emission than in solution. Absorption (Fig. S5†) and emission
spectra of naphthalene diimides were recorded in the solid state.
Solid films were obtained by spin coating the solution of corre-
sponding molecule in chloroform and the emission and excita-
tion spectra are given in Fig. 2. The emission spectrum of NDI-E
showed a monomeric emission peak at 455 nm and aggregate
Fig. 1 (a) UV-Vis absorption and (b) emission spectra of 0.1 OD solu-
tions of naphthalene diimides in chloroform.
corresponding to the 0–0, 0–1 and 0–2 vibronic transitions of
a S0–S1 transition with the intensity of the 0–0 transition being
the maximum.37 No differentiation of ester versus amide linkage
among the molecules was observed in the absorption spectra in
chloroform as solvent.
Fig. 1b shows the fluorescence emission spectra of all the
naphthalene diimide molecules recorded for 0.1 OD solution in
chloroform. The emission spectrum was the mirror image of the
absorption spectrum with the emission maximum centered at 408
nm and other peaks at 387 nm, 432 nm etc.37,38 The quantum
Fig. 2 Solid-state emission (lex ¼ 360 nm) and excitation (lem ¼ 455 nm)
spectra of naphthalene diimides spin coated from chloroform solution.
Table 1 Absorption and emission data of naphthalene diimides
Chloroform
Film
a
lmax abs.
(nm)
b
lmax emission
(nm)
lmax abs.
(nm)
Optical band
gapd
Epc (V)
average
LUMO
(eV)
HOMO
(eV)
lb emission (nm)
fFL
c
NDI-molecule
NDI-E
NDI-A
NDI-E3
NDI-A3
380
380
380
380
387, 408, 432
0.016
0.007
0.003
0.002
368
366
365
362
455 (530, 564)
460 (560)
438 (550)
3.0
2.9
3.0
3.0
ꢁ0.79
ꢁ0.76
ꢁ0.77
ꢁ0.76
ꢁ3.66
ꢁ3.69
ꢁ3.68
ꢁ3.69
ꢁ6.66
ꢁ6.67
ꢁ6.70
ꢁ6.69
387, 408, 432 (552)
387, 408, 432
387, 408, 432 (552)
437 (565)
a
Values corresponds to 0–0 transition. b Excitation at 360 nm and values reported in bracket are for aggregate emission. c The fluorescence quantum
yields were obtained upon excitation at 360 nm and were measured using quinine sulphate as standard. d Calculated from the absorption edge in the solid
state. Epc average: average of first and second reduction peak; (E1/2onset FC ¼ 0.35 V).
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 15235–15246 | 15239

View Online
emission peaks were observed at ꢂ528 nm and ꢂ560 nm (a
similar observation was seen with film drop cast from o-DCB
also). The excitation spectra collected at 530 nm, 570 nm and at
the monomer emission of 455 nm were all identical to the
absorption spectra indicating that the emitting and absorbing
species were identical. NDI-A showed mostly red shifted aggre-
gate emission at ꢂ565 nm with only minor contribution from
monomer emission in the 400–500 nm region. A similar obser-
vation was reported for the carboxyl functionalized unsymmet-
rical NDI derivative, where a broad, featureless red shifted
emission with large stock shift occurred at 545 nm, which was
attributed to the emission from the pre-assembled hydrogen
bond mediated J-aggregate.41 As observed in solution, in solid
state also strong quenching of fluorescence was observed for
NDI-E3 and NDI-A3 (the data have been scaled by a factor of 10
for sake of comparison). The aggregate emission at ꢂ565 was
also observed for NDI-A3, but with less intensity.
Bulk organization
The NDI derivatives were examined for mesogenic characteris-
tics using DSC, PLM and variable temperature WXRD. The
four NDI derivatives of the ester as well as amide series were
found to be thermally stable to 400 ꢃC using thermogravimetric
analysis (TGA, Fig. S7†) and their 10% wt loss temperature is
given in Table S1†. The thermal phase transition temperature
observed from DSC along with the enthalpy of transition is also
given in Table S1†. The amide molecules had higher melting
temperatures compared to the ester molecules due to the
increased rigidity afforded by the hydrogen bonding interaction.
The ester and amide molecules without any terminal alkyl
substitution i.e. NDI-E and NDI-A showed only a melting
transition in their DSC thermogram (Fig. S8†). Upon intro-
duction of the three flexible dodecyloxy units at the terminal
phenyl group, multiple transitions appeared in the DSC ther-
mogram in addition to a drastic reduction in the melting tran-
sitions for the NDI-E3 and NDI-A3 molecules. Fig. 3 shows the
second heating and first cooling cycles in the DSC thermograms
of NDI-E3 and NDI-A3. The PLM (polarized light microscope)
images showing a spherulitic crystalline phase collected at 65 ꢃC
and 132 ꢃC respectively for the two samples NDI-E3 and NDI-A3
are also given in Fig. 3. To understand the nature of the observed
phases, variable temperature WXRD measurements were
collected for both NDI-E3 and NDI-A3. Fig. 4 shows the WXRD
data collected between 2q ¼ 3 to 35ꢃ at various temperature
intervals while cooling from the isotropic melt at 110 ꢃC for NDI-
E3. The small angle region from 2q ¼ 2–5ꢃ was scanned in
a separate experiment and the inset in the figure shows the small
angle data at 2q ¼ 1.91ꢃ for the sample at room temperature
(25 ꢃC). The room temperature WXRD pattern was indexed via
rigorous iteration method (DICVOL43) to a monoclinic cell of
The cyclic voltammetry studies of the NDI derivatives were
done using a three electrode system in dichloromethane (DCM)
solvent
and
tetrabutylammonium
hexafluorophosphate
(TBAPF6 0.1 M–DCM) used as supporting electrolyte. A
representative cyclic voltammogram for NDI-E3 and NDI-A3
was shown in Fig. S6† and the electrochemical data are given in
Table 1. Two reversible reduction peaks corresponding to
formation of radical anion and dianions were observed. The
reduction potential peak values were similar for all the four
molecules demonstrating that they had similar molecular elec-
tronic structures and the electronic properties were insensitive
to the nature of the imide substitution. The LUMO energy
levels of the all NDI derivatives were estimated based on the
average onset value of the first and second reduction peaks and
reference energy level of ferrocene (4.8 V below the vacuum
level) according to ELUMO (eV) ¼ ꢁe ꢀ (Ered onset + 4.8)
below the vacuum level.42 The HOMO levels were calculated
based on the optical band gap obtained from the solid state
absorption onset measurements. The facile reduction, along
with ease of solution processability, makes these NDI deriva-
tives promising candidates for use as n-type materials in
organic field effect transistors.
ꢀ
ꢀ
ꢀ
dimensions a ¼ 47.386 (27) A; b ¼ 3.307 (2) A; c ¼ 24.799 (15) A;
and b ¼ 104.13 (8)ꢃ. The Bragg reflection at 2q ¼ 1.91ꢃ corre-
ꢀ
sponding to d-spacing of 46.06 A was indexed as (100) and hence
assigned as the molecular length peak, and a peak at 2q ¼ 26.94ꢃ
ꢀ
(3.306 A) was indexed as the (010). The molecular length
obtained from the Chem draw 3D energy minimized structure
ꢀ
was 54.06 A. The shorter length observed by XRD indicated
Fig. 3 DSC heating and cooling curves for NDI-E3 and NDI-A3 and PLM images at 65 ꢃC and 132 ꢃC respectively for NDI-E3 (top) and NDI-A3
(bottom).
15240 | J. Mater. Chem., 2012, 22, 15235–15246
This journal is ª The Royal Society of Chemistry 2012

View Online
Fig. 5 Powder WXRD data for NDI-E (top) and NDI-A (bottom).
was indexed to a monoclinic cell with parameters a ¼ 26.592 (21)
ꢃ
ꢀ
ꢀ
ꢀ
A, b ¼ 4.687 (2) A, c ¼ 19.973 (9) A and b ¼ 94.56 (3) . The
ꢀ
length of ꢂ26 A along the ‘a’ axis of the cell was equal to the
ꢃ
ꢀ
observed d-spacing at 26.20 A (2q ¼ 3.36 ) in XRD corre-
sponding to the (100) reflection. No other peaks were observed in
the low angle region <2q ¼ 3ꢃ. Systematic absence evaluation
pointed to a definite lack of screw axis when compared to NDI-A
and the space group was tentatively assigned as P2/m since no
extinction conditions could be observed. A simple model
building exercise based on molecular length comparison to the
unit cell length led to some interesting observations. In the case
Fig. 4 Variable temperature X-ray diffraction pattern of (top) NDI-E3
and (bottom) NDI-A3. Inset shows the peak at low angle (2q ¼ 0.5 to 5ꢃ).
considerable interdigitation of the terminal alkyl chains. In the
isotropic phase at 110 ꢃC, no peaks were observed (Fig. 4), at 65
ꢃC (Fig. 4) sharp peaks were observed both in the small angle and
wide angle region indicating a crystalline phase. Fig. 4 (bottom)
shows the variable temperature WXRD pattern for NDI-A3
taken from the isotropic melt at 150 ꢃC to room temperature
(25 ꢃC). This could also be indexed to a monoclinic cell, a ¼
ꢀ
of NDI-E, the ‘‘b’’ axis length of ꢂ4.68 A restricted any other
orientation than a parallel face to face arrangement for the
naphthalene rings occupying the corners of the unit cell.
However, this distance is longer than the p–p interstack distance
in the range 3.3 to 3.5 A reported for similar systems.4,44 On the
ꢀ
ꢀ
ꢀ
ꢀ
48.040 (36) A, b ¼ 3.831 (3) A, c ¼ 24.846 (18) A and b ¼ 104.09
(9)ꢃ. The clearly distinguishing feature between the XRD
patterns of NDI-E3 and NDI-A3 was the reduced crystallinity in
the latter. NDI-A3 had very few sharp peaks in the entire 2q
range of 3–35ꢃ. Another very important difference compared to
other hand, fluorescence studies in the solid state showed an
absence of quenching of fluorescence for NDI-E (Fig. 2) indi-
cating the absence of face-to-face p stacking interaction in these
molecules. To account for these observations from WXRD as
well as solid state emission, an arrangement with the end phenyl
rings of two naphthalene molecules interdigitated with each
other and the naphthalene rings occupying the unit cell corners
ꢃ
ꢀ
NDI-E3 was that the amide the peak around 2q ¼ 23.37 (3.80 A)
was very sharp and intense and matched with the unit cell length
along the b axis (010), which could be assigned to the p–p-
stacking of NDI-core.4,41,44 Moreover, annealing increased the
intensity of this peak much more than for the ester (Fig. S9†).
Thus, the important observations made by analyzing the
PWXRD pattern of both NDI-E3 and NDI-A3 were (a) NDI-E3
was more crystalline compared to NDI-A3 and (b) the p stacking
ꢀ
was arrived at leading to an a axis length of 26.592 A which
matches with the indexed unit cell. In comparison to this struc-
ture, NDI-A exhibited an elongated a axis; however, the molec-
ular interactions can be summarized to be more or less similar
since the screw axis effects a face centering in this system
ꢀ
reducing the distance between two naphthalene centers to ꢂ23 A
ꢀ
along the b axis was much shorter for NDI-E3 at 3.307 A
ꢀ
(this distance also matched with observed d200 spacing of 22.13 A
compared to 3.831 A for NDI-A3.
at 2q ¼ 3.99ꢃ in XRD). Density Functional Theory (DFT) was
used to obtain energy minimized structures for both NDI-E and
NDI-A by using the turbomole suite of programs.45 Geometry
optimizations were performed using the B–P 86 functional.46,47
Fig. 6a compares the energy minimized structures of NDI-E and
NDI-A with the molecules oriented along the NDI aromatic core.
The NDI core was out of plane with respect to the hexyl and
phenyl planes; however, unlike NDI-E, the amide and phenyl
rings of NDI-A were not totally perpendicular but it was twisted
and had a dihedral angle of ꢂ26ꢃ. This twisting, which occurred
to accommodate the restraining hydrogen bonding, also resulted
ꢀ
The powder XRD pattern of NDI-E and NDI-A showed that
they were more crystalline compared to NDI-E3 and NDI-A3.
Fig. 5 shows the powder X-ray diffraction pattern of NDI-E and
NDI-A. NDI-A was indexed to a monoclinic cell with cell
ꢀ
ꢀ
ꢀ
parameters a ¼ 45.10 (4) A; b ¼ 7.708 (11) A; c ¼ 9.200 (7) A and
b ¼ 99.48 (5)ꢃ. NDI-A had a more layer ordered structure
compared to NDI-E with the peaks indexed as (200), (400), (600)
and (800) along with other higher order (hk1) planes. A careful
observation of systematic absences in the (h00) and (0k0) planes
indicated a plausible non-standard space group of P21/a. NDI-E
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 15235–15246 | 15241

View Online
to the carbonyl of the naphthalene core. However, clearly
differentiating the two (ester vs. amide) was the peak at 1634
cmꢁ1 in the NDI-A (circled peak) which was assigned as the
ꢃ
hydrogen bonded carbonyl stretch. At 250 C the peak at 1634
cmꢁ1 vanished completely and the peak at 1656 cmꢁ1 also shifted
to higher frequency 1667 cmꢁ1. In the NH stretching region also,
the hydrogen bonded NH stretch at 3304 cmꢁ1 reduced in
intensity and shifted to 3327 cmꢁ1 and a new broad peak corre-
sponding to a free NH stretch appeared at 3673 cmꢁ1 (circled
peak). This shift was clearly reversible with the hydrogen bonded
carbonyl and amide peaks shifting to their original values upon
cooling.
Device characteristics
Bottom-gate top contact structures were fabricated for field
effect transistor studies. The fabrication procedure adopted was
as described in the Experimental section. Except NDI-A, all the
other molecules were spin coated from a solution of concentra-
tion 25 mg mlꢁ1 to obtain films of ꢂ80 to 100 nm thickness. In the
case of NDI-A, spin coating could not be done due to poor
solubility in o-DCB and the films could only be drop cast from o-
DCB. This method resulted in a thicker (ꢂ0.4 mm) and more
uniform film. The molecules NDI-E3 and NDI-A3 were heated
above their melting transition temperature 100 ꢃC and 150 ꢃC
respectively and then cooled to room temperature to retain them
in the spherulitic crystalline phase. In the case of NDI-E and
NDI-A, where the melting transition temperature was very high,
the samples were annealed at 150 ꢃC and 200 ꢃC respectively and
then cooled to room temperature to get the optimum device
performance. Fig. 8 shows the typical OFET output and trans-
conductance characteristics of NDI-E (a) and NDI-A (b)
respectively (NDI-E3 and NDI-A3, Fig. S10†) and values of
electron mobility with on/off ratio are reported in Table 2. The
devices showed typical n-type characteristics with a clear tran-
sition from linear to saturation behavior. The performance
parameters were extracted in the saturation regime from the
transconductance characteristics curves by using the following
equation.
Fig. 6 (a) DFT energy optimized structures of NDI-E and NDI-A. (b)
Schematic representation of molecular packing in NDI-E and NDI-A.
in interdigitation to a larger extent in NDI-A. This was reflected
ꢀ
in the shorter d-spacing value of 22.13 A for NDI-A compared to
ꢀ
26.20 A for NDI-E, which is shown schematically in Fig. 6b.
To demonstrate the existence of hydrogen bonding in the solid
state, variable temperature FTIR studies were carried out
(Fig. 7). FTIR spectra of both NDI-E and NDI-A were recorded
in the powder form (pellet made from KBr ground with sample).
Fig. 7 shows the expanded NH stretching region (4000–3300
cmꢁ1) and the carbonyl stretching frequency in the range 1700–
1600 cmꢁ1 at two diffeꢃrent temperatures (at 30 ꢃC and at 250 ꢃC)
for NDI-A and at 30 C for NDI-E. The carbonyl group of the
naphthalene imide has a stretching frequency at 1700 and 1660
cmꢁ1 whereas the carbonyl group of the amide has the amide I
stretching frequency in the range 1650–1690 cmꢁ1. At 30 ꢃC,
NDI-A showed an NH stretching vibration at 3304 cmꢁ1 and
three peaks were observed at 1704, 1656 and 1634 cmꢁ1 in the
carbonyl stretching region. The NDI-E molecule, which clearly
did not have an amide I stretch, also had a peak at 1659 cmꢁ1 due
Ids ¼ (mFET WC/2L) (Vg ꢁ Vth)2
where Ids is the drain current, W and L are, respectively, the
channel width and length, C is the capacitance per unit area of
the gate insulator layer, and Vg and Vth are the gate voltage and
the threshold voltage, respectively. The mFET values of the NDI
molecules were in the order 5.7 ꢀ 10ꢁ3 cm2 Vꢁ1 sꢁ1 (for NDI-A3)
to 2 ꢀ 10ꢁ2 cm2 Vꢁ1 sꢁ1 (for NDI-E) with the NDI-E exhibiting
the most sizable performance parameters. The amide molecules
exhibited relatively lower mobilities. The lower on/off ratio
observed in NDI-A could be attributed to higher off currents
from the residual bulk conductance of the thicker films. Table 2
also compares the mobility values of NDI-E3 and NDI-A3 films
which were not heated to the isotropic phase and then cooled but
only subjected to annealing at temperatures lower than their
melting transitions. The mobility values were found to be one
order of magnitude lower for these low temperature annealed
samples which were not in the spherulitic crystalline phase. The
interdigitation of the terminal alkyl chains in NDI-E3 and
Fig. 7 Expanded FTIR spectra of NDI-E at 30 ꢃC and NDI-A at 30 ꢃC
and 250 ꢃC showing the disappearance of the hydrogen bonded carbonyl
stretch ꢂ1634 cmꢁ1
.
15242 | J. Mater. Chem., 2012, 22, 15235–15246
This journal is ª The Royal Society of Chemistry 2012

View Online
Fig. 8 Typical output and transconductance curves for NDI-E (top) and NDI-A (bottom) from the transistor with W ꢂ 2 mm and L ꢂ 60 to 100 mm
respectively and BCB as the dielectric with C0 ꢂ 4 nF cmꢁ2. Inset shows the schematic of the device.
NDI-A3 resulted in a stacked columnar organization with cofa-
cial alignment along the column,48,49 as indicated by the XRD
Table 2 OFET Characteristics of the naphthalene diimidesa
data. The quenching of fluorescence in the solid state also indi-
cated a cofacial alignment in these 3,4,5-tridodecyloxy termi-
nated NDI derivatives. Therefore, higher mobilities were
observed in the samples annealed at higher temperature
compared to that annealed at low temperatures. However, the
charge carrier mobility in the spherulitic phase of NDI-E3 and
NDI-A3 was still one order lower compared to that obtained for
their non-alkyl chain terminated analogues NDI-E and NDI-A.
This could be attributed to the decreased electron density and the
associated defects because of the alkyl chains in the spherulite
systems.10
mavg
Molecule (cm2 Vꢁ1 sꢁ1
mmax
(cm2 Vꢁ1 sꢁ1
Ion
Ioff
/
Tannealing
(ꢃC)
)
)
Vth (V)
NDI-E
NDI-A
NDI-E3
(a) 2 ꢀ 10ꢁ2
(a) 6 ꢀ 10ꢁ2
>104 ꢁ5 to 15 150
(b) 4.3 ꢀ 10ꢁ4 (b) 5 ꢀ 10ꢁ4
(a) 9.5 ꢀ 10ꢁ3 (a) 1.7 ꢀ 10ꢁ2 >102 ꢁ20 to 20 200
(b) 2.6 ꢀ 10ꢁ4 (b) 4.5 ꢀ 10ꢁ4
(a) 8.5 ꢀ 10ꢁ3 (a) 4 ꢀ 10ꢁ2
(b) 1.5 ꢀ 10ꢁ4 (b) 5.1 ꢀ 10ꢁ4
>104 ꢁ10 to 20 100
>103 ꢁ15 to 30 150
50–60
NDI-A3 (a) 5.7 ꢀ 10ꢁ3 (a) 1 ꢀ 10ꢁ2
(b) 1.4 ꢀ 10ꢁ4 (b) 4.1 ꢀ 10ꢁ4
(a) After annealing. (b) Before annealing.
100
a
The effects of annealing on the thin film morphology of NDI-E
and NDI-A were studied using tapping mode AFM images of
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 15235–15246 | 15243

View Online
thin films deposited on glass. Fig. 9 (top) shows the tapping mode
AFM images of thin films of NDI-E and NDI-A spin coated from
o-DCB both at room temperature (25 ꢃC) and after annealing to
150 ꢃC and 200 ꢃC respectively50 and then cooled to 25 ꢃC. Fig. 9a
and b shows the height and amplitude image of NDI-E at 25 ꢃC
respectively. After heat treatment the sample became smoother
with improved grain size (Fig. 9c and d – the height and ampli-
tude image of NDI-E after annealing respectively). Similarly, the
images on the right side of Fig. 9 correspond to thin films of NDI-
A before (a and b) and after (c and d) annealing. It was observed
that annealing helped improve the overall smooth morphology
of the films.
the preferred orientation in the lamellar direction.29 Due to the
poor solubility of NDI-A, only a very thin film could be obtained
by spin coating from hot o-DCB. Lamellar ordering up to three
orders only could be observed due to the extremely low thickness
of the film (the amorphous hump ꢂ2q ¼ 20ꢃ being diffraction
ꢀ
from glass). However, the lower d200 value (21.62 A) of NDI-A
ꢀ
compared to the molecular length of 33.91 A (from DFT) indi-
cated a tilted orientation on the substrate.6,23,24 In most bottom
gate OFET configurations the preferred orientation of p-
conjugated cores should be approximately perpendicular to the
dielectric surface (edge-on arrangement) for efficient charge
transport from source to drain.52,53 The tilt angle with respect to
the substrate normal was given by cosꢁ1(21/33.91) ¼ 50ꢃ for NDI-
A, and cosꢁ1(26/33.91) ¼ 40ꢃ for NDI-E.2,54 The different dihe-
dral twisting angle, obtained for the energy minimized structure
from DFT, though not important in solution, becomes an
important parameter while considering the orientation on the
substrate surface. In the thin film, this distortion of NDI-A from
planarity resulted in lower observed crystallinity and charge
carrier mobility as compared to NDI-E.11 Another factor
favoring the improved mobility in NDI-E compared to NDI-A
was the increased flexibility of the molecule as a whole due to
absence of the highly restricting directional hydrogen bonding
interaction which gave it more freedom to pack in a more favored
bulk crystalline packing which facilitated electron transport.
NDI-E had isotropic electronic charge transport owing to its
flexible ordered structure. The critical role of restrictive interac-
tions leading to reduced dimensionality on the charge carrier
mobility was shown in the case of C60 derivatives recently. The
group of Dario M. Bassani et al. reported a comparative study of
charge carrier mobility of C60 with that of a derivatized C60
capable of organizing into infinite hydrogen bonded tapes
oriented along one of the crystal axes.30 The charge carrier
mobility of the self-assembled fullerene ribbons was found to be
two-orders of magnitude lower than that of C60 and this was
attributed to the reduced dimensionality of the fullerene tapes as
a wire-like semiconductor in contrast to C60 as a bulk semi-
conductor. We speculate a similar reasoning of restricted
dimensionality of transport for NDI-A which could explain its
lower mobility compared to NDI-E. For instance, the more
soluble NDI-A3 molecule formed gels in non-polar solvents like
methyl cyclohexane (MCH) in contrast to NDI-E3 which
remained crystalline. Fig. 10 compares the TEM images of NDI-
E3 and NDI-A3, respectively, drop cast from MCH solution. It
can be seen that NDI-A3 formed a continuous network of highly
Thin film X-ray diffraction could be used as a tool to under-
stand out of plane ordering of layered films of crystalline small
molecules.12,51 Fig. 9 (bottom) shows the thin film (spin coated
from o-DCB) WꢃXRD data of NDI-E and NDI-A annealed at
ꢃ
ꢃ
150 C and 200 C respectively and then cooled to 25 C. The
percentage crystallinity of a thin film of NDI-E was much higher
compared to the powder with a very high intense (100) peak.
Lamellar ordering up to 6th order could be observed, which was
consistent with edge-on orientation on the substrate.18,19 The thin
film XRD was similar to the powder data (Fig. S11†) indicating
that the crystal structure was retained in the thin film also with
Fig. 9 AFM images of NDI-E and NDI-A (top) taken on glass slide (a)
height (b) amplitude before heating and (c) height and (d) amplitude after
thermal annealing and (bottom) thin film WXRD data for NDI-E and
NDI-A.
Fig. 10 TEM (transmission electron microscope) images for NDI-E3
and NDI-A3 drop casted from 1 ꢀ 10ꢁ4 M solution of MCH.
15244 | J. Mater. Chem., 2012, 22, 15235–15246
This journal is ª The Royal Society of Chemistry 2012

View Online
entangled nanofibers. Although this morphology of NDI-A3 in
MCH cannot be directly correlated with the thin film
morphology in the OFET devices, it could be used to illustrate
the reduced dimensionality of the hydrogen bonded molecules
NDI-A and NDI-A3 compared with their non-hydrogen bonded
ester analogues. Since the overall device performances of the
ester derivatives were better compared to the amides, it could be
concluded that hydrogen bonding in the amides proved to be
more restrictive compared to the more crystalline ester mole-
cules, which had more freedom to realign for efficient charge
transport.
References
1 A. Babel and S. A. Jenekhe, J. Am. Chem. Soc., 2003, 125, 13656.
2 B. A. Jones, A. Facchetti, T. J. Marks and M. R. Wasielewski, Chem.
Mater., 2007, 19, 2703.
3 B. A. Jones, M. J. Ahrens, M.-H. Yoon, A. Facchetti, T. J. Marks and
M. R. Wasielewski, Angew. Chem., Int. Ed., 2004, 43, 6363.
4 B. J. Jung, J. Sun, T. Lee, A. Sarjeant and H. E. Katz, Chem. Mater.,
2009, 21, 94.
5 C. Piliego, D. Jarzab, G. Gigli, Z. Chen, A. Facchetti and M. A. Loi,
Adv. Mater., 2009, 21, 1573.
6 K. C. See, C. Landis, A. Sarjeant and H. E. Katz, Chem. Mater., 2008,
20, 3609.
7 R. Schmidt, M. M. Ling, J. H. Oh, M. Winkler, M. Koenemann,
€
Z. Bao and F. Wurthner, Adv. Mater., 2007, 19, 3692.
8 S. Tatemichi, M. Ichikawa, T. Koyama and Y. Taniguchi, Appl. Phys.
Lett., 2006, 89, 112108.
9 T. D. Anthopoulos, B. Singh, N. Marjanovic, N. S. Sariciftci,
Conclusions
The present work described a full account of the synthesis,
characterization and molecular/thin film properties of two series
of naphthalene diimide (NDI) derivatives – one series had an
amide group linking a terminal phenyl or 3,4,5-tridodecyloxy
phenyl unit to a hexyl spacer substituted NDI core, while in the
other series the linkage was through an ester unit. The amide
linked NDI derivatives formed strong hydrogen bonded aggre-
gates with pronounced red shifted emission both in solution and
solid state. All the four NDI derivatives exhibited good n-type
charge transport characteristics in an OFET device in the bottom
gate top contact geometry with values ranging from 2 ꢀ 10ꢁ2 cm2
Vꢁ1 sꢁ1 to 5.7 ꢀ 10ꢁ3 cm2 Vꢁ1 sꢁ1. The mobility values for the NDI
ester and amide derivatives without terminal alkyl substitution
were one order higher in magnitude compared to their terminal
alkyl substituted counterparts. This was expected, since the alkyl
chains formed an insulating barrier for efficient charge transport
in the NDI-E3 and NDI-A3 derivatives although they also
facilitated formation of cofacially p stacked columns by inter-
digitation. The comparison of mobility values of the ester versus
amides revealed some interesting observations. Contrary to
expectations, the hydrogen bonding in the amide derivatives did
not facilitate better electron mobility. The ester derivatives
exhibited better device performance compared to the amides.
Analyzing the packing in these molecules based on DFT energy
minimized structures and power XRD data, especially in the case
of NDI-E and NDI-A, the NDI amide derivative showed a dihe-
dral twist of the terminal phenyl unit with respect to the naphthyl
aromatic core with a consequent shortening of the molecular
length packing along the a axis. This small difference in the
dihedral twist angle reduced the crystallinity and resulted in
a more tilted orientation of the NDI-A molecules on the substrate
surface. On the whole, the present studies with NDI derivatives
thus demonstrated that the restrictive directional hydrogen
bonding interaction is not so favorable for efficient charge
transport compared to the freedom in charge transport pathways
offered by a three dimensional crystalline organization.
€
A. M. Ramil, H. Sitter, M. Colle and D. M. De Leeuw, Appl. Phys.
Lett., 2006, 89, 213504.
10 T. Vehoff, B. Baumeier, A. Troisi and D. Andrienko, J. Am. Chem.
Soc., 2010, 132, 11702.
11 R. P. Ortiz, H. Herrera, R. Blanco, H. Huang, A. Facchetti,
T. J. Marks, Y. Zheng and J. L. Segura, J. Am. Chem. Soc., 2010,
132, 8440.
12 B. A. Jones, A. Facchetti, M. R. Wasielewski and T. J. Marks, Adv.
Funct. Mater., 2008, 18, 1329.
13 D. Braga and G. Horowitz, Adv. Mater., 2009, 221, 1473.
14 T. Sakanoue and H. Sirringhaus, Nat. Mater., 2010, 9, 736.
15 M. Mas-Torrent and C. Rovira, Chem. Rev., 2011, 111, 4833.
16 M. L. Tang and Z. Bao, Chem. Mater., 2011, 23, 446.
17 E. Ahmed, G. Ren, F. S. Kim, E. C. Hollenbeck and S. A. Jenekhe,
Chem. Mater., 2011, 23, 4563.
18 H. E. Katz, A. J. Lovinger, J. Johnson, C. Kloc, T. Siegrist, W. Li,
Y.-Y. Lin and A. Dodabalapur, Nature, 2000, 404, 478.
19 H. Z. Chen, M. M. Ling, X. Mo, M. M. Shi, M. Wang and Z. Bao,
Chem. Mater., 2007, 19, 816.
20 L. Tan, Y. Guo, G. Zhang, Y. Yang, D. Zhang, G. Yu, W. Xu and
Y. Liu, J. Mater. Chem., 2011, 21, 18042.
21 M. M. Durban, P. D. Kazarinoff and C. K. Luscombe,
Macromolecules, 2010, 43, 6348.
22 M. M. Durban, P. D. Kazarinoff, Y. Segawa and C. K. Luscombe,
Macromolecules, 2011, 44, 4721.
23 P. Gawrys, D. Boudinet, A. Kornet, D. Djurado, S. Pouget, V. Jean-
Marie, M. Zagorska and A. Pron, J. Mater. Chem., 2010, 20, 1913.
24 P. Gawrys, D. Djurado, J. Rimarcik, A. Kornet, D. Boudinet,
V. Jean-Marie, V. Lukes, I. Wielgus, M. Zagorska and A. Pron, J.
Phys. Chem. B, 2010, 114, 1803.
25 P. Deng, Y. Yan, S.-D. Wang and Q. Zhang, Chem. Commun., 2012,
48, 2591.
26 R. T. Weitz, K. Amsharov, U. Zschieschang, E. B. Villas,
D. K. Goswami, M. Burghard, H. Dosch, M. Jansen, K. Kern and
H. Klauk, J. Am. Chem. Soc., 2008, 130, 4637.
27 X. Zhan, Z. Tan, B. Domercq, A. Zesheng, X. Zhang, S. Barlow,
Y. Li, D. Zhu, B. Kippelen and S. R. Marder, J. Am. Chem. Soc.,
2007, 129, 7246.
28 Y. Hu, X. Gao, C.-A. Di, X. Yang, F. Zang, Y. Liu, H. Li and
D. Zhu, Chem. Mater., 2011, 23, 1204.
29 D. Shukla, S. F. Nelson, D. C. Freeman, M. Rajeswaran,
W. G. Ahearn, D. M. Meyer and J. T. Carey, Chem. Mater., 2008,
20, 7486.
30 C.-C. Chu, R. Guillaume, D. Ray, A. D. Guerzo, B. Kauffmann,
G. Wantz, L. Hirsch and D. M. Bassani, J. Am. Chem. Soc., 2010,
132, 12717.
31 D. Yokoyama, S. Hishahiro, Y. Furukawa, C. Adachi and J. Kido,
Adv. Funct. Mater., 2011, 21, 1375.
32 H. E. Katz, J. Otsuki, K. Yamazaki, A. Suka, T. Takido,
A. J. Lovinger and K. Raghavachari, Chem. Lett., 2003, 508.
Acknowledgements
NBK and SPS thank CSIR for SRF fellowship. We thank the
DST project GAP 277826 and network project NWP0054 for
financial support. The authors thank Mr Pandiaraj S. and Pooja
NCL-Pune, for TEM and AFM measurements, respectively.
Manoj Mane’s help with the DFT calculations is also
acknowledged.
€
€
€
33 M. Gsanger, J. H. Oh, M. Konemann, H. W. Hoffken, A.-M. Krause,
€
Z. Bao and F. Wurthner, Angew. Chem., Int. Ed., 2010, 49, 740.
34 T. Seki, Y. Maruya, K. Nakayama, T. Karatsu, A. Kitamura and
S. Yagai, Chem. Commun., 2011, 47, 12447.
35 Y. Miyake, P. Hu, K.-Q. Zhao, H. Monobe, A. Fujii, M. Ozaki and
Y. Shimizu, Mol. Cryst. Liq. Cryst., 2010, 525, 97.
36 M.-X. Zhang and G.-J. Zhao, ChemSusChem, 2012, 5, 879.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 15235–15246 | 15245

View Online
€
H. Weiss, TURBOMOLE; (Version 5.3), Universitat Karlsruhe,
Karlsruhe, Germany, 2000.
37 T. C. Barros, S. Brochsztain, V. G. Toscano, P. B. Filho and
M. J. Politi, J. Photochem. Photobiol., A, 1997, 111, 97.
38 M. E. Ozer, D. Uzun, I. Elci, H. Icil and M. Demuth, Photochem.
Photobiol. Sci., 2003, 2, 218.
46 A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38, 3098.
47 J. P. Perdew, Phys. Rev. B, 1986, 33, 8822.
€
39 F. Wurthner, C. Thalacker, S. Diele and C. Tschierske, Chem.–Eur.
J., 2001, 7, 2245.
48 A. Wicklein, M.-A. Muth and M. Thelakkat, J. Mater. Chem., 2010,
20, 8646.
49 Z. Chen, V. Stepanenko, V. Dehm, P. Prins, L. D. A. Siebbeles,
40 B. Jancy and S. K. Asha, Chem. Mater., 2008, 20, 169.
41 M. R. Molla and S. Ghosh, Chem.–Eur. J., 2012, 18, 1290.
42 N. B. Kolhe, S. K. Asha, S. P. Senanayak and K. S. Narayan, J. Phys.
Chem. B, 2010, 114, 16694.
43 A. Boultif and D. Louer, J. Appl. Crystallogr., 1991, 24, 987.
44 H. Shao, M. Gao, S. H. Kim, C. P. Jaroniec and J. R. Parquette,
Chem.–Eur. J., 2011, 17, 12882.
€
J. Seibt, P. Marquetand, V. Engel and F. Wurthner, Chem.–Eur. J.,
2007, 13, 436.
50 Y.-L. Lee, H.-L. Hsu, S.-Y. Chen and T.-R. Yew, J. Phys. Chem. C,
2008, 112, 1694.
51 B. A. Jones, A. Facchetti, M. R. Wasielewski and T. J. Marks, J. Am.
Chem. Soc., 2007, 129, 15259.
€
45 R. Ahlrichs, M. Bar, H. P. Baron, R. Bauernschmitt, S. Bocker,
M. Ehrig, K. Eichkorn, S. Elliott, F. Furche, F. Haase, M. Haser,
€
52 G. Horowitz, J. Mater. Res., 2004, 19, 1946.
53 Y. Sun, Y. Liu and D. Zhu, J. Mater. Chem., 2005, 15, 53.
54 A. Facchetti, M. Mushrush, M. Yoon, G. R. Hutchison,
M. A. Ratner and T. J. Marks, J. Am. Chem. Soc., 2004, 126,
13859.
€
€
H. Horn, C. Huber, U. Huniar, M. Kattannek, C. Kolmel,
€
€
M. Kollwitz, K. May, C. Ochsenfeld, H. Ohm, A. Schafer,
U. Schneider, O. Treutler, M. von Arnim, F. Weigend, P. Weis and
15246 | J. Mater. Chem., 2012, 22, 15235–15246
This journal is ª The Royal Society of Chemistry 2012