Combinatorial methods for the optimization of the vapor deposition of polyimide monomers and their polymerization

Article abstract of DOI:10.1039/b606091a

A combinatorial study on the synthesis and in situ orientation of thin films of aromatic polyimides on different aligning surfaces was carried out. Monomer and polyimide libraries prepared by using a combinatorial approach consisting of step gradients wit

Full text of DOI:10.1039/b606091a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Combinatorial methods for the optimization of the vapor deposition of
polyimide monomers and their polymerization
Christian Neuber, Markus Ba¨te, Reiner Giesa and Hans-Werner Schmidt*
Received 28th April 2006, Accepted 23rd June 2006
First published as an Advance Article on the web 21st July 2006
DOI: 10.1039/b606091a
A combinatorial study on the synthesis and in situ orientation of thin films of aromatic polyimides
on different aligning surfaces was carried out. Monomer and polyimide libraries prepared by using
a combinatorial approach consisting of step gradients with different thicknesses, sectors with
different compositions, and sectors with or without alignment layers were investigated. As aligning
surfaces, friction deposited layers of PTFE, rubbed polyimide films, and highly oriented polyimide
layers prepared by a shearing technique were used. In addition to 3,39,4,49-biphenyldianhydride
(BPDA), only para-linked dianhydrides and diamines with different aspect ratios were utilized.
Vapor deposition was performed first with individual monomers, then sequentially or by
coevaporation of two or more monomers. By using this combinatorial approach, the monomer and
polyimide orientation can be optimized by variation of monomer composition, film thickness, and
deposition sequence. The resulting films were characterized regarding their thickness and the
achieved degree of orientation by polarized FTIR and UV–vis spectroscopy. The highest degree of
orientation indicated by a dichroic ratio of nine to ten was observed for 4,40-diamino-p-terphenyl
on PTFE surfaces. For other monomers, dichroic ratios of around two were determined
irrespective of the aligning surface, indicating low orientation. It was found that coevaporation of
both monomers yields isotropic polyimide films in good quality, whereas sequential evaporation of
monomer pairs yields polyimide films with anisotropic properties.
(PMDA) and 4,49-diaminodiphenyl ether (ODA). VDP was
employed to synthesize a variety of other polyimides6–8
Introduction
using silylated monomers,9 or for fluorinated polyimides.10,11
The combinatorial approach is a very fast and efficient method
for screening new materials and material combinations, and
Comparing polyimide films prepared by VDP to solvent cast
for optimizing device configurations with the advantage of
films, VDP films showed a smooth surface, improved dielectric
preparing internal references under identical conditions.1,2 In
properties, higher resistance to hygroscopic stresses, improved
this paper we describe the utilization of a combinatorial
stiffness, and a higher density indicating more compactly
packed molecules.8 Polyimides are particularly suitable for
approach for the optimization of the fabrication of polyimide
films via sequential vapor deposition (VD) and vapor deposi-
tion polymerization (VDP) of monomer pairs. The optimiza-
tion of this process requires the control of many experimental
parameters, such as the nature and composition of monomers,
evaporation rate, layer thickness, aligning surface, sequential
or codeposition, and imidization conditions, which can be
largely facilitated by combinatorial methods.
VDP experiments since they are formed in a polyaddition–
condensation sequence and their monomers are frequently
purified by sublimation indicating high thermal stability.
Highly ordered organic thin films are of great interest in the
electronic, optic, and display industries due to their enhanced
properties for various applications.12–14 The orientation of
different molecules can be achieved by self-assembly or by
external energetic or mechanical forces, such as highly
electrical magnetic fields, shear forces, or alignment layers.
For example, buffed polyimide films with an optimized surface
orientation are state-of-the-art in the LC-display techno-
logy.15–17 Here the intermolecular interaction between the
polymer chains of the alignment layer and the LC molecules
are presumably responsible for the alignment. Therefore the
rubbing-induced alignment can be compared to the intrinsic
orientation obtained from bulk-drawn polymers.18,19 Further,
sheared precursor polymers and the corresponding polyimides
also show intrinsic bulk orientation20 and thus they should be
alternatively useable as polyimide alignment layers. A com-
pletely different aligning surface is produced by the friction
deposition of poly(tetrafluoroethylene) (PTFE). Electron-
diffraction studies of the friction-transferred PTFE layers
The advantages of using the vapor deposition process are
a solvent free process, no contamination by dust and other
particles, a smooth surface layer, and a homogeneous, easily
adjustable layer thickness.3 With this technique, thin films of
aromatic polyimides can be prepared, which basically cannot
be obtained via solution processing methods due to the limited
solubility of monomers and the resulting polymers. Polyimide
thin films chemically identical to Kapton1 were first
synthesized by VDP independently by the research groups of
Iijima4 and Salem;5 both evaporated pyromellitic dianhydride
Makromolekulare Chemie I und Bayreuther Institut fu¨r
Makromoleku¨lforschung (BIMF), Universita¨t Bayreuth, Bayreuth,
Germany. E-mail: hans-werner.schmidt@uni-bayreuth.de;
Fax: +49921 55 3206; Tel: +49921 55 3200
3466 | J. Mater. Chem., 2006, 16, 3466–3477
This journal is ß The Royal Society of Chemistry 2006

View Article Online
showed that the chain axis of the PTFE helixes is oriented
parallel to the substrate surface and extended along the sliding
direction.21 The obtained single-crystal-like structure of the
PTFE layers over a large surface (several cm2) aligns low and
high molecular weight liquid crystalline and crystalline
materials from melt, solution, and by epitaxial growth from
vapor.21–24 VDP was further used to fabricate thin films with
intrinsic molecular orientation parallel or perpendicular to the
substrate, which can be controlled by the nature of the
substrate and the deposition conditions. Generally, molecules
orient parallel to the surface at lower substrate temperatures
and perpendicular at higher.25
(0.2 mm PTFE filter) evenly over the substrate. The final layer
thickness can be adjusted by using different blade gap sizes and
solution concentrations.
Monomer synthesis
1,4-Phenylene-N,N9-bis(pyromellitic dianhydride)diimide (6).
10.00 g (45.85 mmol) PMDA (4) were dissolved in a mixture of
500 mL NMP and 100 mL pyridine at 70 uC. While stirring,
1.24 g (11.46 mmol) PDA 5 in 20 mL of this NMP–pyridine
mixture were added dropwise. Then the solution was stirred
for 6 h, cooled, and the product was precipitated in 2 L
toluene. The dianhydride-endcapped monomer 6 was filtered
and purified by sublimation.
As the VDP process and the orientation of polyimides is
complex and affected by many different parameters, the
advantage of an efficient screening method is obvious.
Single, one-by-one experiments are not able to consider the
influence of several complex interacting processing parameters.
Therefore we developed a combinatorial experiment in order
to address the influence of the various parameters. In this
contribution, the evaporation and orientation characteristics
of fully aromatic diamino and dianhydride monomers for the
formation of polyimides on different aligning surfaces are
described. A combinatorial strategy was developed employing
substrates consisting of different aligning surfaces on which
polyimide monomers can be deposited in a step-thickness
gradient. The resulting libraries consist of sectors of deposited
monomers on different aligning surfaces and at different layer
thicknesses. In addition, these sectors can be compared to a
reference area, all prepared in one single experiment under
identical conditions.
Yield: 4.06 g (71%). FT-IR (KBr); v˜ [cm21]: 3076, 1866,
1
1785, 1724, 1522, 1377, 796. H-NMR (DMSO-d6); d [ppm]:
8.17 (b, 4H), 7.64 (b, 4H).
1,4-Diaminophenyl-N,N9-2,5-diethylbis(ethoxycarbonyl)tere-
phthalic acid diamide (9). 2.17 g (15.72 mmol) 4-nitroaniline (7)
were placed in a flame-dried three-neck flask. 50 mL NMP
were transferred into the flask and the mixture was mechani-
cally stirred at RT until all solids dissolved. The solution was
cooled to 210 uC and 1 mL (7.86 mmol) TMSCl was injected.
Then 2.73 g (7.86 mmol) 8 were added at once under nitrogen.
The reaction mixture was stirred for 3 h at 0 uC and then
slowly warmed to RT and kept at this temperature for 12 h.
The resulting solution was precipitated in 1 L water under
vigorous stirring in a blender, and the precipitate was washed
twice with 100 mL ethanol and dried under vacuum at 60 uC.
The reduction to 9 was performed in an autoclave at 4 bar
H2 in N,N-dimethylacetamide catalyzed by 0.10 g Pd/C
(10 wt% Pd).
Experimental
Yield: 3.20 g (74%). FT-IR (KBr); v˜ [cm21]: 3461, 3412,
3366, 3329, 2926, 2870, 1771, 1718, 1633, 1506, 824. 1H-NMR
(DMSO-d6); d [ppm]: 8.15 (s, 2H), 8.12 (dd, 8H), 4.25 (q, 4H),
1.15 (t, 6H).
Materials
3,39,4,49-Biphenyldianhydride, BPDA, (1), pyromellitic dia-
nhydride, PMDA, (4) (both Chriskev Comp. US), p-phenyle-
nediamine, PDA, (5) (Aldrich), and 4,40-diamino-p-terphenyl
(3) (Lancaster) were purified by sublimation. Perylene-
3,4,9,10-tetracarboxylic dianhydride (11) (Lancaster) was dried
under vacuum at 250 uC for 6 h. 4,49-Diamino-1,19-binaphthyl,
naphthidine (2)26 and 2,5-diethylbis(ethoxycarbonyl)tere-
phthaloyl acid (8)27 were synthesized according to literature
procedures. N-Methyl-2-pyrrolidinone (NMP, water content
,0.005%), pyridine, and chlorotrimethylsilane (TMSCl) were
purchased from Aldrich in Sure Seal1 bottles and used
without further treatment. m-Cresol was refluxed and distilled
from CaH2 and stored under dry argon atmosphere. All other
solvents and chemicals were commercial products and used
as received.
N,N9-Bis-(1,4-diaminophenyl)-3,4,9,10-perylene bis(diimide)
(12). 1.03 g (2.63 mmol) perylene dianhydride 11 and 2.84 g
(26.25 mmol) PDA 5 were refluxed with a small amount of
Zn(Ac)2?2H2O as catalyst in 50 mL m-cresol. After 12 h the
mixture was cooled to room temperature, the solvent was
removed, and the obtained perylene diamine 12 was purified
by sublimation.
Yield: 0.52 g (35%). FT-IR (KBr); v˜ [cm21]: 3170, 3063,
1687, 1589, 1359, 1274. UV(solid film): 495sh, 500, 560 nm.
Alignment layer preparation
Substrates were cleaned successively in chloroform and
isopropyl alcohol for 10 min in an ultrasonic bath, kept in
methyl ethyl ketone steam for about 3 h, dried in vacuum at
50 uC, and finally treated in an oxygen plasma etcher (Anatech
Plasma Series DAIWA DP810 Digital SWR Power Meter,
1.5 6 1022 bar at 70 W for 10 min) immediately before use.
For doctor blading of solutions, a custom-made machine
(BASF AG) was used. A glass substrate was mounted on a
heated metal plate at 70 uC. A motor-driven doctor blade
(Erichsen, barrel shape: 5 mm gap) moved a filtered solution
A commercial buffing machine LCBM6 (Optron Systems Inc.)
with a translating substrate holder, a vacuum chuck, and a
rotating roller (diameter 70 mm) covered with a velvet cloth
was used. The roller-to-stage distance was adjusted with two
micrometer set-screws. The buffing process was carried out at
RT, the revolution of the roller was 600 rpm, the pile
impression was 0.8 mm, and the substrate was moved in one
continuous motion at 1 mm s21
Surface polyimide aligning substrates were obtained by
doctor blading precursor polymer solutions on glass
.
a
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 3466–3477 | 3467

View Article Online
substrate. The obtained films were first soft-baked at 100 uC
for 1 h and then thermally converted into the polyimide under
N2 atmosphere at a hot stage. The samples were kept for 1 h
at 100 uC, 200 uC, and 300 uC, ramped at 10 K min21. The
polyimide-covered substrates were buffed at RT; only two-
thirds of each substrate were buffed to create a reference sector
with no buffing treatment. Bulk oriented polyimide substrates
were fabricated via a shearing process using a Tribotrak1
machine from DACA Instr., USA, with a modified setup
described in detail elsewhere.20 Lyotropic solutions of a
poly(amic ethyl ester) were sheared with a sequence of four
1.5 mm diameter glass rods mounted on a holder at 25 mm s21
at 100 uC and 10 kg. To minimize the evaporation of NMP, the
lyotropic solution was covered with an additional glass plate
during the shearing process. After shearing, the PAE films
were thermally imidized as described above.
slowly and so the substrate was stored in inert atmosphere
between the different characterization methods so that
oxidation can be neglected.
Characterization
FTIR spectra were recorded in transmission mode (16 scans;
resolution: 4 cm21) using a BioRad Digilab FTS-40 and KBr
pellets. Polarized transmission FTIR spectra were recorded on
silicon wafers in combination with a KRS 5 wire grid polarizer
from BioRad. For measurement with parallel and perpendi-
cular polarized light, the sample was kept in the same position
but the polarizer was rotated by 90u.
1H-NMR Spectra were recorded using a Bruker AC 250
spectrometer (250 MHz) in deuterated dimethyl sulfoxide
(DMSO-d6). UV–vis absorption spectra were acquired on a
spectral photometer (Hitachi U-3000). Polarized absorption
spectra were measured on quartz substrates, placing one Glan–
Taylor parallel-oriented polarizer in the sample optical path
and in the reference path, respectively. For measuring parallel
and perpendicular absorption spectra, the polarizers were
rotated 90u, keeping the sample in the same position, and
baselines for each particular aligning surface were recorded.
Polarized optical microscopy was performed on a Leitz
Laborlux 12-Pol. For layer thickness measurements, a surface
profilometer (Dektak 3030 ST) was used. The film was
scratched with a scalpel and the resulting gap depth denoted
as the layer thickness of the film. Settings at medium speed
combined with medium resolution and a stylus force of
1?1025 N result in an accuracy of ¡ 5 nm. Thermal imidiza-
tion of evaporated layers was performed on a nitrogen-flushed
hot plate controlled by a Eurotherm 903P. First, the hot plate
was heated at 1 K min21 to 100 uC; at this temperature the
substrate was tempered for 6 h. The temperature was increased
in steps at 1 K min21 to 150 uC, kept for 12 h, raised again at
1 K min21 to 300 uC, kept for 12 h, and finally raised at
1 K min21 to 350 uC and annealed for 12 h.
For friction deposition of PTFE on quartz glass substrates
or silicon wafers, the same Tribotrak machine was employed,
consisting of a translating stage and a pivoting PTFE rod
holder. The speed of the translating stage can be varied
between 0.01 and 30 mm s21, the sample stage and the PTFE
holder can be heated individually between 25 uC and 350 uC.
On top of the PTFE holder contact, a force is applied by
weights in the range of 2.5 to 10 kg. Only two-thirds of each
substrate was covered with PTFE alignment layer to create a
one-third reference sector without a layer. To evaluate the
effect of temperature on the quality of the PTFE aligning
surfaces, the substrate and PTFE holder temperature were set
to 150 and 300 uC in combination with weights of 5 and 10 kg.
Combinatorial vapor deposition
Each substrate was mounted in a rotatable substrate holder
above a mask which is located next to the motor-driven
shutter. About 300 mg of each monomer were placed in the
sources using crucibles. The vacuum chamber (Balzers PLS
500) was evacuated to 1025–1026 mbar. First, a constant
evaporation rate was set manually by increasing the tempera-
ture of the sources; then, control was transferred to a computer
program to achieve reproducible and constant evaporation
rates. The closed loop control input was the frequency signal
of quartz crystals placed near the source. Here, evaporation
rates were about ten times higher than detected with additional
quartz sensors near the substrate. Thus, slight fluctuations of
the evaporation rate were corrected by the hardware, resulting
in a very homogenous layer formation on the substrate. After a
few minutes of equilibration, the shutter was opened and the
deposition on the substrate proceeded. A step motor moved
the shutter in five steps at 530 seconds per step creating a five-
step gradient. During normal operation, the mask is missing
and the shutter creates the steps. Optionally, different masks
can be inserted which cover part of the substrate; for example,
by using three different one-third-open masks, which are open
on the upper, middle, and lower third, respectively, three
different sections can be created on the substrate. In this way,
the numbers of sectors can be additionally increased. At the
end of the experiment, the shutter was closed, the source power
switched off, and the vacuum chamber ventilated with air. The
oxidation of the evaporated diamines in air proceeds very
Results and discussion
Monomer deposition
Monomer selection. Vapor deposition requires that mono-
mers are thermally stable during evaporation. Additionally,
vapor deposition polymerization needs polymerizable mono-
mers in the solid state, which once deposited show reduced
sublimation at the subsequent thermal polymerization. More
details about the setup and combinatorial technique are
published elsewhere.1 Low molecular weight monomers are
known to desorb during the evaporation process, particularly
at higher substrate temperatures; later during the course of the
thermal imidization, unreacted monomers and oligomers can
also sublime.10 Monomers for aromatic polyimides fulfil these
requirements since they are thermally stable, are commonly
purified by sublimation, and most are commercially available.
First, monomers for aromatic rod-like polyimides were
selected which possess an intrinsic structural anisotropy,
meaning their shape is more elongated than spherical (Fig. 1).
This anisotropy should be helpful for achieving a high
orientation of the deposited monomers on aligning surfaces.
3468 | J. Mater. Chem., 2006, 16, 3466–3477
This journal is ß The Royal Society of Chemistry 2006

View Article Online
monomer 1029 (Fig. 2b). The condensation conditions were
similar to a typical low temperature synthesis of aromatic
polyamides in NMP,30 and an intermediate compound with
nitro-protected amino groups was obtained. The autoclave
reduction with hydrogen rendered the diamino monomer 9 in
good yields. This precursor was directly used for further
experiments since thermal imidization will occur during
evaporation in the crucible and only the imidized molecule
10 is deposited.
Fig. 1 Commercially available polyimide monomers used for vapor
deposition.
3,39,4,49-Biphenyldianhydride (1, BPDA) was chosen in
addition to para-substituted diamines such as 4,49-diamino-
1,19-binaphthyl (2) and 4,40-diamino-p-terphenyl (3).
In order to introduce a UV–vis active chromophore for
polarized UV measurements even on polyimide surfaces,
whose absorption overlaps with the one of the monomers
presented so far, a perylene unit was functionalized with
diamine groups. For the synthesis (Fig. 2c), the perylene
dianhydride 11 and an excess of PDA 5 were refluxed with
Zn(Ac)2?2H2O as catalyst in m-cresol. The resulting deep
crimson monomer 12 was purified by sublimation.
Several series of combinatorial experiments were performed:
evaporation on Type 1 substrates for evaluating the optimal
processing conditions for PTFE aligning surfaces, Type 2
substrates were used to elucidate the difference between PTFE
and polyimide aligning surfaces, and Type 3 and 4 substrates
were used for investigating the formation of polyimides during
sequential and simultaneous evaporation of monomer pairs,
respectively. To further increase the aspect ratio and molecular
weight of monomers, other compounds were developed.
Monomer 6 was synthesized by endcapping p-phenylenedi-
amine, PDA, (4) with two PMDA moieties as shown in Fig. 2a.
Under typical conditions of a one-step polycondensation for
polyimides (i.e. stoichiometric ratio, m-cresol, and reflux),26,28
and in combination with a large excess (10 : 1) of PMDA, an
oligomer mixture was always obtained. However, by using a
5 : 1 mixture of NMP and pyridine as solvent and an excess of
PMDA, monomer 6 was yielded almost quantitatively at a
moderate reaction temperature of 70 uC. To completely avoid
oligomer formation, 4-nitroaniline (7) and bis(ethoxycarbonyl)
terephthalic dichloride (8) were used for the synthesis of
Combinatorial design for selecting the optimal PTFE aligning
surface. To study the deposition of monomers and their
orientation on PTFE aligning surfaces prepared at different
conditions, substrate Type 1 (Fig. 3) was developed. Row A
consists of Si-wafers glued onto the upper quartz glass slide
without any alignment layer for FTIR measurements. Row B
is a friction-deposited PTFE alignment layer prepared at
150 uC and 10 kg. The center quartz slide is coated with two
PTFE alignment layers prepared at 300 uC/10 kg (row C) and
300 uC/5 kg (row E). Row D serves as a reference row without
any aligning surface. The bottom quartz slide is coated with a
PTFE alignment layer prepared at 150 uC/5 kg in row F and
Si-wafers with PTFE alignment layers (300 uC/10 kg) in row G.
Fig. 2 a) Synthesis of the anhydride-endcapped monomer 6 by using an optimized process and stoichiometry (see text). b) Synthesis of the amino-
endcapped monomer 10. The diamino diester 9 was thermally imidized during the evaporation process, yielding 10. c) Synthesis of diamino-
functionalized perylene compound 12.
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 3466–3477 | 3469

View Article Online
Fig. 3 Combinatorial substrate Type 1 (76 6 75 mm) for investigat-
ing the orientation behavior of polyimide monomers. The dashed lines
indicate the use of three standard 76 6 25 mm glass slides. The
substrates are quartz glass with PTFE aligning surfaces (row B, C, E,
and F) and 9 6 9 mm Si-wafers (row A and G) for UV–vis and FTIR
spectroscopy, respectively. The PTFE aligning surfaces have been
prepared under different conditions regarding the temperature and
pressure. Row D serves as a reference row without any alignment layer.
Fig. 5 Layer thickness dependence of naphthidine (2) at three
evaporation rates on substrate Type 1. The layer thickness increases
linearly over the five-step thickness gradient independently of the
evaporation rate. Each step was deposited for 530 seconds.
Naphthidine (2) was evaporated at three different evaporation
rates onto different sectors of the same substrate using three
different sequentially positioned one-third-open masks, mean-
ing they are open on the upper, middle, and lower third,
respectively.
For each rate, a five-step thickness gradient was fabricated
by moving the shutter in five steps to the left. Fig. 5 and Table 1
show that the layer thickness increases linearly over the five
steps independently of the evaporation rate which was
calculated from the slope of the linear fit at 0.07, 0.14, and
0.21 nm s21. Furthermore, the average step thickness of
monomer 2 evaporated at different rates is listed: at a set
evaporation rate of 0.21 nm s21, the produced step thickness
increases from 110 nm (step 1) to 530 nm (step 5), whereas at a
rate of 0.07 nm s21, a thickness of 45 nm to 180 nm was
determined. At equal evaporation conditions, BPDA (1)
and 4,40-diamino-p-terphenyl (3) show a very similar linear
increase in layer thickness. Since it is known that the
orientation behavior of evaporated molecules is influenced
by the evaporation rate,31 all monomers were evaporated at
very low evaporation rates between 0.03 to 0.06 nm s21. In all
cases, a linear relationship of the step thickness and evapora-
tion rate could be established permitting a prediction of layer
thickness as a function of time. However, this parameter must
be determined for each monomer in separate experiments.
Further, it was found that the substrate shows homogeneous
thickness for each step independently of substrate type
(glass, Si-wafer) or aligning surfaces. Hence, the thickness of
deposited layers is solely governed by the evaporation rate and
Fig. 4 Sketch of monomer deposition as
a five-step thickness
gradient on a Type 1 substrate (76 6 75 mm). A step is generated
by moving the shutter below the substrate to the left in steps, so that on
column 5 the thickest layer is deposited. The rows A to G identify
different substrates such as quartz glass or Si-wafers, columns 1 to 5
steps with different monomer layer thickness. Hence, a combinatorial
library with 35 sectors is obtained.
Thus, a combinatorial substrate is obtained with four different
alignment layers (B, C, E and F) on quartz glass, two layers on
Si-wafers (A and G), and an uncovered quartz glass sector for
reference (D).
Monomers 1 to 3 were evaporated as a five-step thickness
gradient as sketched in Fig. 4. These gradients were generated
by moving the shutter and no mask in five steps; thus columns
1 to 5 specify different layer thickness of the deposited
monomers and column 5 bears the thickest layer.
First, a correlation of evaporation rate and layer thickness
was established. The layer thickness was determined with the
Dektak profile scanner as described in the Experimental part.
Table 1 Layer thickness dependence of deposited monomers at different evaporation rates
Average step thicknessb/nm
Measureda
Monomer
Substrate type
evap. rate/nm s21
Step 1 (530 s)
Step 2 (1060 s)
Step 3 (1590 s)
Step 4 (2120 s)
Step5 (2650 s)
1
2
2
2
3
6
10
12
a
1
1
1
1
1
2
2
2
0.20
0.07
0.14
0.21
0.06
0.06
0.03
0.05
110
45
80
110
40
30
15
30
b
230
70
150
240
70
63
30
50
345
120
240
340
110
90
430
150
300
420
150
120
60
530
180
360
530
170
153
80
50
65
95
120
Slope of the best fit regression line as shown in Fig. 5. Experimentally determined by mechanical profile measurements.
3470 | J. Mater. Chem., 2006, 16, 3466–3477
This journal is ß The Royal Society of Chemistry 2006

View Article Online
overlaps with the wavelength of the evaporated molecules.
Consequently, polarized UV–vis measurements only make
sense on PTFE aligning surfaces. To overcome this restriction,
monomer 12 was synthesized with a perylene chromophore
which possesses absorption maxima at 500 and 560 nm,
distinct from those of polyimides. UV–vis spectra were
measured on quartz glass rows B, C, E, and F. For FTIR
spectroscopy, the Si-wafers of rows G and K were removed
from the substrate and placed in the beam. The calculated
dichroic ratios (DR = intensity of the transmitted polarized
light parallel to the director of the alignment layer divided
by the intensity of the transmitted perpendicular light) on
substrates Type 1 and 2 are summarized in Table 2.
Fig. 6 Combinatorial substrate Type 2 (76 6 75 mm) for investigat-
ing the deposition behavior of polyimide monomers on different
aligning surfaces. The quartz glass substrate is coated with sheared
(row I) and rubbed (row L) polyimide alignment layers and with a
friction-deposited PTFE layer prepared at 300 uC/10 kg (row J). Row
H, with uncoated glass, and row M, covered with an unoriented
polyimide film, serve as references, whereas row K contains Si-wafers
with PTFE alignment layers (also 300 uC/10 kg) for FTIR spectro-
scopic investigations.
BPDA (1) and naphthidine (2) possess an absorption
maximum at 300 and 350 nm, respectively, but polarized
UV–vis spectroscopy was not able to detect any orientation in
the deposited layers indicated by a DRUV of 1 in Table 2.
Polarized FTIR spectra of 1 and 2 on Type 1 substrates show
no indication of orientation. The orientation of monomer 3 on
PTFE aligning surfaces was characterized by polarized UV–vis
spectroscopy between 240 and 550 nm on rows B, C, E, and F.
The spectra show a broad maximum at 340 nm and an
orientation with a distinct dependence of the measured DRUV
on layer thickness. The DRUV increases up to ten at a layer
thickness of about 100 nm (step 2–3) and decreases at thicker
layers independently of the surfaces underneath. Polarized
FTIR spectroscopy was also employed to characterize the
orientation of the deposited monomer 3. A characteristic mode
is the symmetric vibration of the 1,4-substituted benzene unit
not by the substrate surfaces or alignment material, meaning
the results can be adopted for other substrates.
Combinatorial design for comparing polyimide and PTFE
aligning surfaces. On substrate Type 2 (Fig. 6), the orientation
behavior of monomers on three different types of aligning
surfaces, such as PTFE, sheared polyimide, and buffed
polyimide, was studied. In detail, PTFE aligning surfaces at
row K and J, with row K for FTIR characterization, were
prepared at 300 uC/10 kg. The uncoated reference row H is
followed by a polyimide alignment layer (row I) fabricated by
shearing a concentrated polyimide precursor solution.20 The
bottom quartz slide was completely covered by doctor blading
an isotropic poly(amic ethyl ester) solution (solid content
1.5 wt%) in NMP, then dried and thermally imidized. Half of
the slide was rubbed (row L); the other half (row M) serves as a
reference layer. Thus, Type 2 contains five different surfaces:
three different aligning surfaces and two uncovered reference
sectors. Evaporation rates and layer thickness characterization
are included in Table 1.
at 1491 cm21 32
For the determination of the degree of
.
orientation, this symmetric stretching mode, which lies parallel
to the direction of the molecule, was used, and a DRIR of 9 for
sector G5 was calculated at 1491 cm21. Here, only the thickest
layer (sector G5) was investigated since the polarized FTIR
measurements require a layer thickness of above 100 nm for
sufficient absorbance. Summarizing, the investigated prepara-
tion conditions of PTFE aligning surfaces seem to play only a
minor role for the orientation behavior of the investigated
monomers 1 to 3. For further investigations, PTFE aligning
surfaces prepared at 300 uC and with 10 kg were used.
Orientation behavior of the deposited monomers
Polarized UV–vis measurements of monomer 6 on Type 2
substrates in the range of 260 to 550 nm reveal an increase in
absorption only below 350 nm and no distinct maximum. The
degree of orientation for sector J5 (see Fig. 6, PTFE aligning
surface, quartz glass) was determined at 300 nm to DRUV = 2
(Table 2). This indicates only a slight orientation of this
Polarized spectroscopy. Polarized UV–vis and FTIR spectro-
scopy were employed to characterize quantitatively the
achieved orientation of the deposited monomers on top of
aligning surfaces. However, in the investigated range of 240 to
550 nm, the electron absorption of polyimide aligning surfaces
Table 2 Dichroic ratios (DR) of monomers 1 to 12 on PTFE aligning surfaces calculated from polarized UV–vis and FTIR measurements at the
given wavelength or wave number, respectively
Monomer
Substrate type
Sector
Wavelength/nm
DRUV
Sector
Wave numbera/cm21
DRIR
1
2
3
6
10
12
1
1
1
2
2
2
B1
C1
C1
J5
J5
J2
300
350
400
300
330
500
1
1
10
2
2
2
G1
G1
G5
K5
K5
K2
1616
1513
1491
1520
1510
1590
1
1
9
2
2
2
I2
L2
2b
3.2c
a
b
Symmetric stretching mode. Sheared polyimide aligning surface. Rubbed polyimide aligning surface.
c
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 3466–3477 | 3471

View Article Online
rod-like monomer. FTIR spectra of 6 were acquired for
sector K5 (PTFE aligning surface, Si-wafer) and the degree of
orientation was determined using the symmetric stretching
intense birefringence proportional to layer thickness was
observed for 4,40-diamino-p-terphenyl (3), indicating a high
degree of orientation.
mode of the 1,4-substituted phenylene unit at 1520 cm21
,
Selected images of monomer 10 at different sectors of a
Type 2 substrate are shown in Fig. 7. Images I1–I5 show the
birefringence on top of the shear-oriented polyimide surface
next to the uncovered quartz glass. The slight increase in
birefringence is proportional to the layer thickness starting
from sector I1 to I5, while on quartz glass (row H, top upper
half of I1 to I5), no birefringence occurs and thus only
isotropic deposited layers are observed. Another behavior was
observed on rubbed (row L, image L1–L5) and untreated
(row M) polyimide surfaces. Here, 10 crystallizes in small,
homogenously-dispersed crystals, accompanied by an increase
in birefringence proportional to the layer thickness. Lower
birefringence, observed by the polarizer parallel to the rubbing
direction, and higher birefringence at 45u both indicate that
monomer 10 orients parallel to the rubbing direction of the
polyimide surface. Images J1 to J5 depict the same monomer
10 evaporated on a PTFE aligning surface also showing
birefringence intensity proportional to layer thickness.
The color of the deposited steps for perylene monomer 12
is visible with the naked eye and changes proportional to
the layer thickness from green, blue, gray, light red to red
(not shown). The maximal birefringence for this monomer
deposited on PTFE aligning surfaces was observed in sectors
J2 and J3; in this case, a thicker layer did not generate more
intensity. On rubbed and unrubbed polyimide layers, increas-
ing birefringence proportional to layer thickness indicates
orientation on these surfaces which is, however, accompanied
by a slight crystallization of the monomer with increasing
layer thickness.
which oscillates parallel to the direction of the molecule, and a
DRIR of 2 was calculated.
Polarized UV–vis measurements of monomer 10 were
recorded between 550 and 260 nm and showed an absorption
maximum at 330 nm with a shoulder at 300 nm. A DRUV of 2
at 330 nm was calculated for sector J5 (PTFE aligning surface,
quartz glass). The FTIR spectra of 10 show a characteristic
splitting of the asymmetric and symmetric carbonyl-stretching
modes at 1760 cm21/1770 cm21 (shoulder) and at 1730 cm21
/
1700 cm21, respectively. The degree of orientation was
determined using the symmetric stretching mode of the
benzene unit at 1510 cm21, which oscillates parallel to the
direction of the molecule, and a DRIR of 2 was calculated
also for sector K5 (PTFE aligning layer, Si-wafer, Table 2).
Thus, monomers 6 and 10 show a very similar orientation
behavior on PTFE aligning surfaces but the achieved degree of
orientation is low.
UV–vis spectra of monomer 12, which contains the perylene
chromophore, exhibit a shoulder at 495 nm and two maxima at
500 and 566 nm. This distinct bathochromic shift of around
300 nm, compared to the absorption of the other investigated
monomers, permits the investigation with polarized UV–vis
spectroscopy even on polyimide aligning surfaces. The
absorption at 566 nm is typical of the formation of J-type
aggregates, where it is observed that coplanar aggregates of
the dye have parallel transition dipole moments.33 Polarized
UV–vis spectra were recorded between 700 and 300 nm for
row I (sheared polyimide surface, quartz glass) and L (rubbed
polyimide surface, quartz glass). For row I the degree of
orientation was determined to a DRUV of 2, thus in the same
range found for the PTFE aligning surface (row J), while the
rubbed polyimide aligning surface (row L) induces a somewhat
higher DRUV = 3.2 (see Table 2).
Polymerization of vapor deposited monomer pairs
Monomers can be vapor deposited sequentially, meaning one
after another, or simultaneously realized in a coevaporation of
both monomers at the same time. In sequential experiments,
the evaporation monomer sequence was also reversed to
investigate the influence on film thickness and orientation of
the obtained polyimide films. For polymerization, exclusively
PTFE friction deposited layers as an aligning surface were
investigated, since dianhydride monomers 1 and 6 showed only
minor orientation on these surfaces. In all cases the thermal
imidization (see Fig. 8) was accomplished by a thermal
treatment under nitrogen at stepwise increased temperatures
up to 350 uC in combination with slow heating rates of
1 K min21 and long dwell times of 12 h. These optimized
conditions enhance orientation and polymerization, minimize
evaporation, and were found in preliminary experiments.
In summary, on PTFE aligning surfaces, a low DR of two
was determined for monomers 6, 10, and 12, indicating a poor
orientation. Based on these results, we conclude that rod-like
unsubstituted full aromatic monomers have less mobility on
aligning surfaces due to their stiffness and higher molecular
weight. In addition, it seems likely that the atomic grooves
34
˚
between aligned PTFE chains at about 4 A, which account
for the ordered growth of evaporated molecules, do not permit
sufficient reorientation of the monomers.
Polarized optical microscopy. Polarized transmission optical
microscopy is another method to evaluate qualitatively the
orientation of polyimide monomers on top of aligning
surfaces. First, the birefringence on Type 1 substrates was
investigated. Dianhydride BPDA (1) crystallizes in small
crystals on the PTFE aligning surfaces and in large crystals
on quartz glass, and reveals no orientation on the entire
substrate area. Thus the deposition of BPDA is governed by
its fast crystallization regardless of the surface on which it
was deposited. For naphthidine (2), weak birefringence was
observed for thin films only; accordingly, at a step thickness
above 70 nm, no birefringence was visible anymore. The most
Combinatorial design for sequential evaporation on PTFE
aligning surfaces. A combinatorial substrate Type 3 similar to
Type 1, as shown in Fig. 9, was developed to investigate
various aspects of the formation and orientation of aromatic
polyimides prepared by sequential vapor deposition. The
substrate consists of three PTFE alignment layers (row O, P,
and R) deposited on quartz glass substrates at 300 uC and
10 kg. In addition, five Si-wafers for each row (row N, Q,
3472 | J. Mater. Chem., 2006, 16, 3466–3477
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Fig. 7 Polarized light microscopy images of different sectors on a Type 2 substrate library by the vapor deposition of monomer 10 at an
evaporation rate of 0.03 nm s21. The highest intensity of birefringence is observed on the polyimide aligning surfaces in sector I5 and L5. Hence, for
this particular monomer, polyimide layers are more potent aligning surfaces than PTFE layers, but with indications of crystallization.
and S) with identically prepared PTFE aligning surfaces were
mounted on the substrate. Areas without any alignment layers
are found, for instance, between row N and O, and thus can be
compared to films formed on orientating sectors. Polyimide
aligning surfaces were not investigated since monomers 1 and 6
only showed orientation on PTFE aligning surfaces; addition-
ally, diamino functionalized monomers 10 and 12 show
crystallization on polyimide aligning surfaces.
For the sequential evaporation of monomers, the com-
binatorial setup in the vacuum chamber was equipped with a
shutter and a two-thirds-open mask. The first monomer was
evaporated at 0.05 nm s21 in a five-step thickness gradient on
Fig. 9 Combinatorial Type 3 substrate (76 6 75 mm) for investigat-
ing the formation and orientation of aromatic polyimides via
sequentially evaporated monomers. The substrate consists of PTFE
aligning surfaces prepared at 300 uC/10 kg on quartz glass (row O, P
and R) and Si-wafers (row N, Q and S), respectively.
Fig. 8 Thermal imidization of the aromatic dianhydrides 1 and 6 (R)
and the aromatic diamines 2, 3 and 10 (R9) forming the corresponding
polyimide conducted at 100 uC for 6 h, 150 uC/12 h, 300 uC/12 h,
and finally at 350 uC/12 h heated at 1 K min21 under a nitrogen
atmosphere.
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 3466–3477 | 3473

View Article Online
Fig. 10 Preparation of a combinatorial library on a Type 3 substrate (76 6 75 mm) by sequential evaporation of two monomers. A shutter and a
two-thirds-open mask move in five steps to the left creating a five-step thickness gradient. By using this partly open mask and rotating the substrate
180u, the upper and lower third of the substrate are covered by monomer I or II. Only rows P and Q are covered by both monomers with a
thickness step gradient in opposite direction. Rows P and Q are also shown in a side view.
two-thirds of the substrate by moving the shutter and the mask
to the right in five steps, as shown in Fig. 10. Accordingly, step
five has the thickest layer of the first monomer. Then the
substrate was rotated 180u and the shutter and mask returned
to their start position. By repeating the procedure, the second
monomer was evaporated at the same rate also in a five-step
gradient in the opposite direction with respect to the gradient
of the first monomer. Since two-thirds of the substrate were
covered during each move, rows N and O were covered with
the first monomer, rows R and S with the second monomer,
and only rows P and Q were covered with both monomers.
The combinatorial library consists of 30 sectors, 10 sectors
for each monomer (five different thicknesses and two different
rows) and 10 for sequential step gradients in opposite
directions (rows P and Q) plus additional reference areas.
After removing the Si-wafers, areas of uncovered quartz glass
are accessible for further reference if needed.
(,50 nm) were found. Conversely, for step 3, the thickest
polyimide layer was obtained. This demonstrates that the
imidization takes place through a larger part of the film and
diffusion of both monomers penetrating each other is required.
By changing the deposition sequence, meaning evaporating
diamine 10 first and then dianhydride 6 at the same evapora-
tion rate of 0.05 nm s21, very similar results were obtained,
with step 3 possessing the thickest polyimide layer. Sequential
evaporation of other monomer pairs (see Table 3) produced
comparable results and always showed a substantial decrease
of layer thickness after imidization. The location of the sector
with the thickest polyimide film depends on the sequence
of monomer deposition as well as monomer structure.
Theoretically, the thickest polyimide film is to be expected at
strict stoichiometrical conditions. However, identical layer
thicknesses do not assure these conditions since the monomer
layer density, and thus the mass or amount of moles of
deposited material, is different for each individual monomer
and inherently not known.
To investigate the effect of the monomer evaporation
sequence, the dianhydride was deposited directly on top of
the PTFE aligning surface and then covered by the diamine. In
a second experiment the diamine was deposited first followed
by the dianhydride. The layer and polymer film thickness
were mechanically measured for each sector. Fig. 11 shows
the thickness for all steps obtained after first depositing
dianhydride 6. The left bars show individual thicknesses for
monomer 6 and 10 measured in row O and R, respectively. The
middle bars denote the total thickness of row P, which can be
directly compared to the added thickness determined for each
monomer. The right bars show the polyimide film thickness
after imidization (row P).
Further experiments were conducted to incorporate the
perylene diamine as a UV–vis active chromophore. However,
this diamine could not be used alone since its UV–vis
absorption is much too strong. Thus it was deposited as a
Comparing the left bars, the step gradients and the change in
the composition of 6 and 10 are evident. Moving from step 1 to
5, the thickness ratio of monomer 6 increases, with a ratio
close to 1 at steps 2 and 3. The added thickness values of row
O and R (left bar) are in good agreement with the measured
thicknesses of row P formed by the sequential evaporation of
both monomers in an opposite step gradient (middle bar). The
overall thickness of the steps is not constant but increases
slightly with increasing amount of the dianhydride 6 showing
values in the range between 130 and 155 nm. After thermal
imidization, row P (right bars) showed a volume shrinkage
mainly caused by evaporation of excess monomer during
imidization. Consequently, for step 1 and 5 with the largest
differences in composition, the thinnest polyimide layers
Fig. 11 Layer thickness characterization of two sequentially
deposited monomers on substrate Type 3. Left bar: layer thickness
of the reference sectors (row O and R) of each individual monomer; 6
was deposited first. The bar in the center represents the measured
thickness of row P, containing monomer 10 on top of 6 after sequential
evaporation in the form of opposite step gradients. The right bar
reflects the layer thickness of the same step after thermal imidization.
3474 | J. Mater. Chem., 2006, 16, 3466–3477
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Shrinkage caused by the imidization, i.e. loss of water, can
be neglected since the number of benzene units, and thus their
contribution to intensity, remains constant.
Table 3 Layer thickness of selected sectors of the combinatorial
library Type 3 obtained by sequential evaporation before and after
thermal imidization (deposition rate: 0.05 nm s21
)
Best sectora for polyimide
preparation and thickness
(nm) before and after
imidizationb
Polarized UV–vis spectroscopy was also employed to
quantify orientation. For evaporating BPDA (1), which
crystallizes on PTFE aligning surfaces, and the bulky
monomer 2, anisotropy was detected neither for the monomers
nor for the polyimide films, independently of the evaporation
sequence. By evaporating 1 first and then the linear diamine 3,
a DRUV of 2 was determined before and after imidization. But
by reversing the evaporating sequence, the DRUV decreased
from 5 to 2 before and after imidization, respectively. Thus,
the orientation of rigid-rod aromatic polyimides prepared by
vapor deposition onto PTFE aligning surfaces is characterized
by a DRUV of around 2 and it could be shown that during the
thermal cyclization the degree of orientation is not enhanced.
In sector P5 of the sequentially evaporated monomers 10,
12, and 6, UV–vis measurements at 330 nm revealed a DRUV
of 2 independently of the evaporation sequence. After thermal
imidization, this weak anisotropy is maintained, indicated
by the same DRUV value. Consequently, the evaporation
sequence for the rod-like monomers 6, 10, and 12 does
not influence the orientation behavior before and after
imidization.
Monomer Selected monomer sectors
sequence
and thickness/nm
1st–2nd
1–2
2–1
1–3
3–1
6–10
10–12c–6
a
Monomer I Monomer II Sector Before After
O1: 20
O1: 20
O3: 90
O5: 70
O3: 85
—
R1: 100
R1: 90
R3: 65
R5: 50
R3: 50
—
P1
P1
P3
P5
P3
P2
130
110
155
125
145
250
b
100
60
90
40
80
250
Sector with the thickest resulting polyimide layer. For imidization
c
conditions see Experimental part. Thickness: ca. 10 nm.
thin interfacial layer of 10 nm in a sandwich structure between
the diamine 10 (first) and the dianhydride 6 (third). As shown
in Table 3, no shrinkage in sector P2 during thermal imidiza-
tion could be measured for this layer sequence. Before and
after imidization, the thickness was 250 nm in both cases.
Therefore, it was surprisingly found that by adding a thin
perylene layer in between the diamine and the dianhydride, the
shrinkage is dramatically reduced during imidization com-
pared to layers built by two monomers only.
Polarized light microscopy (no micrographs shown) was also
used to characterize the orientation behavior of the polyimide
films. The sequentially evaporated monomers 1 and 2 revealed,
independently of evaporation sequence, a crystalline morpho-
logy on the entire substrate, which is clearly reduced after
imidization due to desorption of monomers. Birefringence was
observed at an angle of 45u (PTFE director to polarizer) for
the sequential evaporation of diamine 3 first, then BPDA (1),
which is slightly decreased after imidization.2 For the reversed
evaporation sequence, only crystalline sectors were observed
before and after imidization. By evaporating diamine 10 first
and then dianhydride 6, the intensity of the birefringence on
sectors Q1 to Q5 is slightly increased proportional to the step
layer thickness. Another behavior is observed for the reversed
evaporation sequence, where a maximum birefringence was
noted for step 3 and a decrease in intensity was observed for
higher step layer thickness. For both evaporation sequences,
the films on top of neat glass surfaces were isotropic and non-
oriented. After imidization, birefringence is clearly lower for
all sectors due to evaporation of monomers during imidiza-
tion. For polyimide films prepared by sequential evaporation
of diamine 10 and dianhydride 6, a maximum of birefringence
was always observed on sector Q3, meaning the monomer
ratio seems to be optimal for this sector, where the thickest
polyimide layer was also found. For the sequential evaporation
of diamine 10 (first), perylene diamine 12 (second), and
dianhydride 6 (third), an isotropic layer on the neat glass
surfaces and a weak birefringence on the sectors with
aligning surfaces were observed, which was slightly stronger
after imidization.
Characterization of the sequentially generated libraries. For
all monomer combinations, the formation of polyimide films
was demonstrated by FTIR spectroscopy. Fig. 12 shows the
FTIR spectra of sector Q3 before (top) and after (bottom)
thermal imidization of the dianhydride 6 on top of diamine 10.
The absence of the N–H stretching modes at 3370 and
3460 cm21 combined with a missing anhydride carbonyl
absorption at 1865 cm21 indicates a high conversion to the
polyimide. However, the obvious decrease in intensity of the
symmetric vibration of the 1,4-substituted benzene unit at
1515 cm21 implies monomer desorption off the substrate
during imidization. This is supported by thickness measure-
ments revealing a thinner polyimide film after imidization
(105 nm) compared to 140 nm of deposited monomers.
Fig. 12 FTIR spectra of sector Q3 before (top) and after (bottom)
thermal imidization of the sequential evaporated monomer 10 (first)
and 6 (second). The layer thickness of 140 nm as deposited is reduced
to 105 nm after thermal imidization. This results in a significantly
lower FTIR absorbance of the imidized layer.
Characterization of libraries generated by coevaporation. Less
complex combinatorial substrates (Type 4) were designed for
the coevaporation of two monomers (Fig. 13). Here, row T
serves as a reference; additionally, one PTFE aligning surface
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 3466–3477 | 3475

View Article Online
thickness of 40 nm (T1), 80 nm (T2), 110 nm (T3), 140 nm
(T4), and 180 nm (T5) was measured before imidization. The
corresponding polyimide films revealed a slightly decreased
layer thickness of 35 nm (T1), 70 nm (T2), 110 nm (T3), 145 nm
(T4), and 175 nm (T5). Similar results were obtained by
coevaporating monomer 1 with 3 and 6 with 10. In these cases,
the observed small shrinkage during imidization is mainly
caused by the loss of water during cyclization and, to a much
lower extent as observed by the sequential experiments, by
monomer desorption.
Fig. 13 Design of combinatorial library Type 4 (76 6 75 mm) for the
coevaporation of two monomers simultaneously. Besides a reference
area (row T), a PTFE aligning surface on quartz glass (row U) and on
Si-wafers (row V) is present. A shutter without a mask moving in five
steps to the left creates a thickness step gradient.
Different stages during the polycondensation and thermal
imidization of coevaporated dianhydride 6 and diamine 10
were monitored by FTIR spectroscopy at different tempera-
tures. The formation of the poly(amic acid) and the progres-
sing imidization as a function of time and temperature have
been documented.2
on a quartz glass substrate (row U) and five Si-wafers with
PTFE aligning surfaces (row V) are included. A shutter was
moved in steps to create a five-step thickness gradient. It was
shown in literature that in order to obtain high molecular
weight polyimide films, the ratio of both monomers must be
optimized and is not necessarily strictly stoichiometric.6 Thus,
coevaporation must proceed at an optimized evaporation
ratio. Since in the sequential experiments described above the
thickest polyimide films were detected for equal monomer
layer thicknesses, it is safe to approximate that an equal
evaporation rate will yield the best polyimide films. On the
contrary, different evaporation rates for each monomer would
render monomer layers with the one monomer in excess, which
was evaporated at a higher rate. Therefore both monomers
were evaporated simultaneously at the same evaporation rate
of 0.05 nm s21. The thermal imidization was accomplished
with the same temperature program as used above.
UV–vis spectroscopy of the coevaporated monomers BPDA
(1) with naphthadine 2, 1 in combination with terphenyl (3),
and dianhydride 6 with diamine 10, and their corresponding
polyimides was also conducted. The investigations of the
orientation behavior show that 1 plus 2 oriented only slightly
and 1 plus 3 oriented with a DRUV of 2 before and after
thermal imidization. A different behavior was observed for the
coevaporation of monomers 6 and 10; here, only isotropic
films on top of and next to PTFE aligning surfaces were
observed.
To finally compare the DRUVs of all investigated mono-
mers, monomer pairs, and resulting polymer layers on PTFE
alignment layers a summary is given in a bar chart in Fig. 14.
The plot shows that the degree of orientation remains low
regardless of the monomer or monomer pairs deposited, except
for monomer 3 and the sequential evaporation 3–1. Thus a
high degree of orientation of both monomers, the dianhydride
and the diamine, is absolutely necessary for the preparation of
highly ordered polyimide films.
Characterization of the sectors was performed by mechani-
cal thickness measurements, polarized light microscopy, as
well as polarized FTIR and UV–vis spectroscopy. Layer
thickness and the degree of orientation can be evaluated as a
function of the chemical nature of the monomers during
coevaporation. For coevaporated monomers 1 and 2, a layer
As a result, the coevaporation of two monomers is the
preferable way to synthesize polyimide thin films as the
desorption of monomer during thermal imidization is
Fig. 14 Comparison of DRUVs of all investigated monomers, monomer pairs, and resulting polymers. The dichroic ratios are calculated at the
corresponding absorption maxima at the optimized layer thickness.
3476 | J. Mater. Chem., 2006, 16, 3466–3477
This journal is ß The Royal Society of Chemistry 2006

View Article Online
negligible. This is presumably due to the formation of CT
complexes. Polyimide films prepared by sequential evapora-
tion show much larger desorption and hence undesired
shrinkage. It should be noted that the obtained thin polyimide
films cannot be prepared by a standard solution polyconden-
sation method due to the insolubility of the poly(amic acid)s
and polyimides.
about two was determined by polarized spectroscopy which
was maintained after imidization.
Acknowledgements
This work was supported by the German Science Foundation
(SFB 481), project A6.
References
Conclusion
1 M. Thelakkat, C. Schmitz, C. Neuber and H.-W. Schmidt,
Macromol. Rapid Commun., 2004, 25, 204.
2 M. Baete, C. Neuber, R. Giesa and H.-W. Schmidt, Macromol.
Rapid Commun., 2004, 25, 371.
3 T. Miyamae, K. Tsukagoshi, O. Matsuoka, S. Yamamoto and
H. Nozoye, Langmuir, 2001, 17, 8125.
4 M. Iijima, Y. Takahashi, K. Inagawa and A. Ito, Shinku, 1985, 28,
437.
5 J. R. Salem, F. O. Sequeda, J. Duran, W. Y. Lee and R. M. Yang,
J. Vac. Sci. Technol., A, 1986, 4, 369.
6 C. C. Roberts, S. A. Letts, M. D. Saculla, E. J. Hsieh and
R. C. Cook, Fusion Technol., 1999, 35, 138.
7 F.-Y. Tsai, E. L. Alfonso, D. R. Harding and S. H. Chen, J. Phys.
D: Appl. Phys., 2001, 34, 3011.
8 J.-H. Jou, C.-L. Cheng, E. C.-L. Jou and A. C.-M. Yang, J. Polym.
Sci., Part B: Polym. Phys., 1996, 34, 2239.
The combinatorial approach is an efficient way to investigate
the complex interplay between the evaporation and orientation
behavior of vapor deposited monomers before, during, and
after thermal imidization. Aromatic rod-like monomers for
polyimides were evaporated in a physical vapor deposition
process on different aligning surfaces. As a substrate, a
combination of three different types of aligning surfaces was
used: friction-deposited layers of PTFE prepared under
different conditions, surface oriented rubbed polyimide films,
and bulk oriented polyimide layers. Also, different substrate
materials, such as quartz glass and Si-wafers for UV–vis
and FTIR spectroscopy, respectively, were employed. This
combinatorial approach permits the study of the influence
of several parameters, such as nature and composition of
monomers, evaporation rate, layer thickness, aligning surface,
and sequential or simultaneous deposition sequence in an
9 M. Iijima, Y. Takahashi, Y. Oishi, M. Kakimoto and Y. Imai,
J. Polym. Sci., Part A: Polym. Chem., 1991, 29, 1723.
10 Y. Y. Maruo, Y. Andoh and S. Sasaki, J. Vac. Sci. Technol., A,
1993, 11, 2590.
11 F.-Y. Tsai, D. R. Harding, S. H. Chen and T. N. Blanton, Polymer,
2003, 44, 995.
evaporation experiment on combinatorial substrates.
A
12 M. Grell and D. D. C. Bradley, Adv. Mater., 1999, 11, 895.
13 K. S. Whitehead, M. Grell, D. D. C. Bradley, M. Jandke and
P. Strohriegl, Appl. Phys. Lett., 2000, 76, 2946.
14 F. C. Grozema, T. J. Savenije, M. J. W. Vermeulen,
L. D. A. Siebbeles, J. M. Warman, A. Meisel, D. Neher,
H.-G. Nothofer and U. Scherf, Adv. Mater., 2001, 13, 1627.
15 N. A. J. M. van Aerle and A. J. W. Tol, Macromolecules, 1994, 27,
6520.
16 Y. B. Kim and B. S. Ban, Liq. Cryst., 1999, 26, 1579.
17 A. Mosley, B. M. Nicholas and P. A. Gass, Displays, 1987, 8, 17.
18 J. M. Geary, J. W. Goodby, A. R. Kmetz and J. S. Patel, J. Appl.
Phys., 1987, 62, 4100.
19 Y. B. Kim and B. S. Ban, Liq. Cryst., 1999, 26, 1579.
20 C. Neuber, R. Giesa and H.-W. Schmidt, Adv. Funct. Mater., 2003,
13, 387.
shutter/mask setup rendering a five-step thickness gradient
controls the thickness of each sector of the library. Aromatic
dianhydrides and para-substituted diamines were used as
monomers. The step thickness gradients of the libraries in
the range of 20 to 530 nm were determined by mechanical
measurements. Dianhydride BPDA showed high crystallinity
independent of the aligning surface without any indication of
orientation. Naphthidine also showed very poor orientation.
On PTFE aligning surfaces, 4,40-diamino-p-terphenyl showed
a dichroic ratio of nine and ten, determined by polarized
spectroscopy. However, the deposition of the monomers
with even higher aspect ratios resulted in a rather low degree
of orientation with a dichroic ratio around two, regardless
of the aligning surface used. Based on these results, we
conclude that rod-like aromatic monomers exhibit low
mobility on aligning surfaces due to their stiffness and high
molecular weight.
21 J. C. Wittmann, S. Meyer, P. Damman, M. Dosiere and
H. W. Schmidt, Polymer, 1998, 39, 3545.
22 J. C. Wittmann and P. Smith, Nature, 1991, 352, 414.
23 R. E. Gill, G. Hadziioannou, P. Lang, F. Garnier and
J. C. Wittmann, Adv. Mater., 1997, 9, 331.
24 P. Damman, M. Dosiere, M. Brunel and J. C. Wittmann, J. Am.
Chem. Soc., 1997, 119, 4633.
Additionally, the formation of polyimides by sequentially
and coevaporated monomers was investigated on com-
binatorial substrates. Polyimides were formed on each sector
and during imidization, a layer shrinkage mainly caused by
desorption of excess monomer was observed. In contrast,
the shrinkage of coevaporated monomers after imidization
was found to be negligible. The orientation behavior of
coevaporated monomers on top of the PTFE aligning surfaces
strongly depends on the molecular structure of the monomers.
The reason might be the formation of monomer charge
transfer complexes and reorientation during the thermally
activated polycondensation step. A rather low dichroic ratio of
25 A. Kubono, H. Higuchi, S. Umemoto and N. Okui, Thin Solid
Films, 1993, 232, 256.
26 R. Giesa, U. Keller, P. Eiselt and H. W. Schmidt, J. Polym. Sci.,
Part A: Polym. Chem., 1993, 31, 141.
27 W. Volksen, Adv. Polym. Sci., 1994, 117, 111.
28 S. Yoshida and A. S. Hay, Macromolecules, 1997, 30, 5979.
29 A. J. McKerrow, M. A. Fox, J. Leu and P. S. Ho, J. Polym. Sci.,
Part A: Polym. Chem., 1997, 35, 319.
30 K. H. Becker and H.-W. Schmidt, Macromolecules, 1992, 25, 6784.
31 A. Kubono and N. Okui, Prog. Polym. Sci., 1994, 19, 389.
32 M. Hesse, H. Meier and B. Zeeh, Spectroscopic Methods in Organic
Chemistry, Georg Thieme Verlag, Stuttgart, New York, 1997.
33 L. B. A. Johansson and H. Langhals, Spectrochim. Acta, Part A,
1991, 47A, 857.
34 T. Tanaka and M. Ishitobi, J. Phys. Chem. B, 2002, 106, 564.
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 3466–3477 | 3477