A conformational study of aromatic imide compounds. Part 1. Compounds containing diphenyl ether and benzophenone moieties

Article abstract of DOI:10.1016/S0022-2860(01)00707-4

An attempt was made to estimate the dihedral angles φ, ψ, ω₁ and ω₂, of bis(4-hydroxyphthalimide)s (BHPI) and bis-(phenylphthalimides)s (BPI) having diphenyl ether or benzophenone linkages at the center of molecules using solid-state ¹³C CP/MAS NMR and ab initio nuclear shielding calculations. The TOSS and TOSS & DD pulse sequences were performed in the NMR measurements to obtain exact chemical shifts of each carbon. Total energies were calculated using the B3LYP/6-31G(d) level of theory, and shielding constants were calculated using the RHF/6-31G(d) level of theory for diphenyl ether and benzophenone by varying the angles of φ, ψ from 0 to 180° at intervals of 10°. The comparison of nuclear shielding calculations between RHF/6-31G(d) and B3LYP/6-311 +G(2d,p) levels of theory indicates that RHF/6-31G(d) has sufficient reliability for calculating shielding constants of diphenyl structures. Since the nuclear shieldings are displaced depending on the dihedral angles, the dihedral angles of BHPIs and BPIs can be estimated by comparing the experimental chemical shifts with the calculated shielding constants. The BHPIs, BPIs and a related polyimide ODPA/ODA were estimated to have symmetric conformations having identical φ and ψ and identical ω₁ and ω₂.

Full text of DOI:10.1016/S0022-2860(01)00707-4

Journal of Molecular Structure 602±603 '2002) 405±416
www.elsevier.com/locate/molstruc
A conformational study of aromatic imide compounds. part 1.
Compounds containing diphenyl ether and benzophenone moietiesq
*
Keitaro Aimi, Toshihiro Fujiwara, Shinji Ando
Department of Organic and Polymeric Materials, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo, 152-8552, Japan
Received 21 March 2001; accepted 2 April 2001
Abstract
An attempt was made to estimate the dihedral angles f, c, v1, and v2, of bis'4-hydroxyphthalimide)s 'BHPI) and bis-
'phenylphthalimides)s 'BPI) having diphenyl ether or benzophenone linkages at the center of molecules using solid-state 13C
CP/MAS NMR and ab initio nuclear shielding calculations. The TOSS and TOSS & DD pulse sequences were performed in the
NMR measurements to obtain exact chemical shifts of each carbon. Total energies were calculated using the B3LYP/6-31G'd)
level of theory, and shielding constants were calculated using the RHF/6-31G'd) level of theory for diphenyl ether and
benzophenone by varying the angles of f, c from 0 to 1808 at intervals of 108. The comparison of nuclear shielding calculations
between RHF/6-31G'd) and B3LYP/6-3111G'2d,p) levels of theory indicates that RHF/6-31G'd) has suf®cient reliability for
calculating shielding constants of diphenyl structures. Since the nuclear shieldings are displaced depending on the dihedral
angles, the dihedral angles of BHPIs and BPIs can be estimated by comparing the experimental chemical shifts with the
calculated shielding constants. The BHPIs, BPIs and a related polyimide ODPA/ODA were estimated to have symmetric
conformations having identical f and c and identical v1 and v2. q 2002 Elsevier Science B.V. All rights reserved.
Keywords: CP/MAS NMR; Shielding constant; Dihedral angle; Diphenyl structure; Polyimide
1. Introduction
regarded as a repeating unit of polyetherimides [2].
The conformation of BHPI is de®ned by four dihedral
angles f, c, v1 and v2 shown in Scheme 1 'Molecular
structures of BHPI, BPI and ODPA/ODA). There are a
few reports that clarify conformations of diphenyl
linkages by experimental methods except for X-ray
diffraction studies of low molecular compounds
[3±65].1,2 Although X-ray diffraction requires single
crystals, it is generally dif®cult to prepare single
crystals of polymers.
Aromatic polyimides and polyetherimides have
been widely used as interlayer dielectrics for IC tips,
high strength ®lms, and high temperature adhesives
due to the high thermal and mechanical stability and
good electric insulating ability [1]. Such mechanical
and physical properties should be closely related to
the conformations and ¯exibility of polymer chains.
Bis'4-hydroxyphthalimide) 'BHPI) that has two imide
rings and a diphenyl '±Ph±X±Ph±) linkage can be
Moreover, most of high-performance aromatic
polymers, such as polycarbonate, polyether-ether-
ketone and polyphenylenesul®de, have diphenyl
linkages in their main chain, and the physical properties
q
Dedicated to Professor Graham A. Webb on the occasion of his
65th birthday.
*Corresponding author. Tel.: 181-3-5734-2137; fax: 181-3-
5734-2889.
1
For diphenyl ether 'DPO) compounds, see Refs. [4±19]
2
E-mail address: sando@polymer.titech.ac.jp 'S. Ando).
For benzophenone 'DPCO) compounds, see Refs. [20±65]
0022-2860/02/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved.
PII: S0022-2860'01)00707-4

406
K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
diphenyl ether calculated using DFT method 'B3LYP/
6-31G'd) level of theory) gives the same accuracy as a
high-level ab initio method 'MP2) [70]. Hence, more-
over, diphenyl linkages in polymer chains in the solid
state do not necessarily have the conformations
predicted for model compounds. In order to estimate
conformations of polymers, experimental procedures
that are applicable to non-crystalline or polycrystalline
samples should be used.
Tonelli et al. have investigated the conformation of
cyclic poly'p-phenylene sul®de), whose dihedral
angles had been determined by X-ray diffraction,
using solid-state NMR and ab initio shielding constant
calculation [71]. They reported that the chemical
shifts of the carbons that are adjacent and ortho to
sul®de atom showed the dependence on dihedral
angles of diphenyl sul®de. The present authors have
reported the dihedral angle 'u) dependence of
chemical shifts of aromatic carbons in biphenyl and
its substituted derivatives in the solid-state [72], in
which the chemical shifts show systematic correlation
with u, when the substituent shielding effects are
removed from the solid chemical shifts by subtracting
the solution chemical shifts, or subtracting the sub-
stituent shielding parameters. The calculated nuclear
shieldings of the carbons that are adjacent to biphenyl
linkage showed good agreement with the reduced
chemical shifts.
Scheme 1.
are related to the ¯exibility of the polymer chain.
Several attempts have been made to predict the stable
conformations of diphenyl structures by molecular
orbital 'MO) calculations [66±70]. Anwer et al. has
predicted the stable conformations of 2,2-diphenyl
propane and six diphenyl molecules by calculating the
conformational energy surfaces using the MM2
molecular mechanics routines [67]. Makowski has
reported the minimum energy conformations of benzo-
phenone and three diphenyl molecules using AM1
semi-empirical method [68]. Straûner has reported the
energy surfaces of diphenylmethane and diphenyl ether
using three semiempirical method 'AM1, MNDO and
PM3) and density functional theory 'DFT) [69].
Although semiempirical methods spend much shorter
computational time than ab initio theories and DFT
methods, the accuracy of the former is much less than
the latter. According to Straûner's report, the surface
shapes obtained from four different methods are similar
to each other, but the conformations of energy minima
are signi®cantly different. Thisindicates thatthe confor-
mations predicted by MO calculations depend on the
theory used. On the other hand, Straûner reported that
the most stable conformation of diphenylmethane and
In this article, we examine the conformations of
BHPIs and BPIs having diphenyl ether and benzo-
phenone moieties using solid-state 13C NMR and ab
initio shielding constant calculations and explored the
possibility to apply this method to estimate conforma-
tion of polyimides.
2. Experimental
2.1. Synthesis
The general formula of BHPI and BPI are shown
in Scheme 1. Two BHPIs th0at have diphenyl linkages
of X ˆ ±O± 'N,N0-'4,4 -oxydiphenylene)-bis-'5-
hydroxy-2-phenyl-isoindole-1,3-dione), BHPI-DPO)
and ±CO± 'N,N0-'4,40-benzophenone)-bis-'5-hydroxy-
2-phenyl-isoindole-1,3-dione), BHPI-DPCO) were
synthesized according to the literature [2]. In order to
exclude the effect of intermolecular hydrogen bonding,

K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
407
two bis'phenylenephthalimide) 'N,N0-'4,40-oxydiphe-
nylene)-bis-'2-phenyl-isoindole-1,3-dione) 'BPI-DPO)
and
N,N0-'4,40-benzophenone)-bis-'2-phenyl-iso-
indole-1,3-dione) 'BPI-DPCO)), which have the same
skeletal structures as those of BHPI but no hydroxyl
groups, were also synthesized. Moreover, ODPA/
ODA polyimide 'Scheme 1) that can be regarded as
the polymer of BHPI-DPO were synthesized to evaluate
the applicability of this method to polymers.
The synthesis of BHPI-DPCO is described as
follows as an example. A mixture containing 10.8 g
of N-methyl-2-pyrrolidinone 'NMP) and 5.4 g of
o-xylene was added to 2.70 g '16.48 mmol) of 4-
hydroxyphthalic anhydride and 1.75 g '8.24 mmol)
of 4,40-diaminobenzophenone. This mixture was
then heated under re¯ux to 1908C for several hours
to remove the water of reaction by azeotropic distilla-
tion. The mixture was made to remain at room
temperature and was then poured into methanol. The
solid product was ®ltered out and rinsed with acetone,
and dried in vacuum at 1008C for 6 h. The obtained
4,40-bis'4-hydroxyphthalimide)benzophenone
Scheme 2.
DEPT sequence was used for selective observation of
protonated carbons. Tetramethylsilane 'TMS) was
used as the internal standard ꢀdc ˆ 0 ppm†:
Solid-state 13C CP/MASNMR spectrawereobserved
with JEOL GSX-270 spectrometer operating at
67.8 MHz. The total suppression of spinning sideband
'TOSS) pulse sequence was used to suppress spinning
sidebands. The pulse sequence that combines TOSS and
the dipolar dephasing 'non-quaternary carbon suppres-
sion) 'TOSS & DD) was alsoused for selective observa-
tion of non-protonated carbons. The CP contact time
was 2 ms, and the recycle delay was 5 s. The chemical
shifts werecalibratedindirectlythroughthe adamantane
peak observed at lower frequency '29.5 ppm relative to
tetramethylsilane 'TMS)).
'BHPI-DPCO) was 3.36 g '80.8% yield). The
corresponding BPIs were synthesized in the same
manner using phthalic anhydride instead of 4-hydro-
xyphthalic anhydride. BHPI-DPO was recrystallized
from acetone, and BHPI-DPCO was recrystallized
from NMP/methanol. BPIs were recrystallized from
chloroform. The powder sample of ODPA/ODA was
synthesized in a similar way to that of BHPIs using
equimolar amounts of 4,40-oxydiphthalic anhydride
and 4,40-diaminodiphenyl ether. On the other hand,
®lms of ODPA/ODA were prepared as follows.
Equimolar amounts of 4,40-diaminodiphenyl ether
and 4,40-oxydiphthalic anhydride were mixed in
N,N-dimethylacetamide 'DMAc) at a weight concen-
tration of 10% and stirred for 1 day at room tempera-
ture. The ODPA/ODA polyamic acid thus prepared
was spin-coated onto a glass plate, dried at 708C for
1 h, and then thermally cured at 1608C for 2 h and
3508C for 1 h.
2.3. MO Calculations
Two dihedral angles, f and c, of diphenyl ether
'DPO) and benzophenone 'DPCO) were varied from 0
to 1808 at 108 intervals 'Scheme 2, Model structure
for MO calculations. X ˆ ±O±'DPO) and ±CO±
'DPCO). The dihedral angles of the shown conforma-
tion are ꢀf; c† ˆ ꢀ08; 08†. Geometrical parameters
'bond lengths and bond angles) were optimized for
each conformation using the B3LYP/6-31G'd) level
of theory, and 13C nuclear shielding constants were
calculated using RHF/6-31G'd) level of theory and
GIAO-CHF method [73]. Cheeseman et al. refer to
this basis combination as the minimum recommended
for NMR properties [74]. The program package of
molecular orbital theory used was gaussian-98
installed in the Computer Center of Tokyo Intitute
of Technology [75].
2.2. NMR measurements
3. Results and discussion
Solution-state 13C NMR spectra were observed with
JEOL GSX-500 spectrometer operating at 125 MHz.
Dimethylsulfoxide 'DMSO-d6) was used as a solvent
for BHPIs, and chloroform-d was used for BPIs. The
3.1. Conformational energy surfaces
Fig. 1 shows the calculated total energy surface of

408
K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
Fig. 1. Conformational energy surface of diphenyl ether 'DPO)
calculated using B3LYP/6-31G'd) level of theory. The dihedral
angles of diphenyl ether structures reported for single crystals are
plotted by ®lled circles within the range of 08 # f # 908 and c $ f
'see text).
Fig. 2. Conformational energy surface of benzophenone 'DPCO)
calculated using B3LYP/6-31G'd) level of theory. The dihedral
angles of benzophenone structures reported for single crystals are
plotted in the same manner of Fig.1.
in good agreement with the results from a high level
calculation of MP2/6-31G'd) '39.48) [70], and the
surface shapes of the two B3LYP levels are almost
identical.
DPO. The energy minima are found at ꢀf; c† ˆ
ꢀ408; 408† and '1408, 1408), and a large minimum
region ',2 kJ/mol) extends from '08, 908) to '908,
08) and from '908, 1808) to '1808, 908). Since the left
and right halves of the energy surface are point-
symmetrical with respect to the pint of '908, 908),
we consider only the left half. The energy difference
between the minimum point and '08, 908) is 0.67 kJ/
mol, and that between the minimum point and '908,
908) is 17.07 kJ/mol. This indicates that the transition
accompanying the rotation of benzene rings between
two energy minima, i.e. '408, 408) and '1408, 1408),
requires an activation energy of 0.79 kJ/mol at '108,
808) and passes the conformation of '08, 908). Anwer
et al. reported a calculated conformational energy
surface of DPO using MM2 [67]. The energy
minimum '388, 388) and the overall shape of the
energy surface given by MM2 are very similar to
those of Fig. 1. According to the calculations using
semiempirical 'AM1) and DFT 'B3LYP/3-21G'd))
methods, the energy minima of DPO are found at f ˆ
c ˆ 358 'AM1), 378 'B3LYP/3-21G'd)) and 408
'B3LYP/6-31G'd)) [69]. In addition, the energy
minima obtained from the B3LYP level surfaces are
Fig. 2 shows the conformation energy surfaces of
DPCO. The energy minimum is found at ꢀf; c† ˆ
ꢀ308; 308†; and the minimum region ',2 kJ/mol),
which is located between '208, 408) and '408, 208),
is narrower than that of DPO. The energy difference
between the minimum point and '08, 908) is 11.50 kJ/
mol, which is signi®cantly higher than that of DPO
'0.79 kJ/mol). In addition, the energy difference
between the minimum point and '908, 908) is
45.57 kJ/mol, which indicates that the conformation
of '908, 908) is very unstable. These results indicate
that benzophenone linkage has very low ¯exibility for
the benzene ring rotation. Anwer et al. has predicted
that the most stable conformation of DPCO is
'308, 308) [67], which agrees well with our result.
Other minimum conformations reported for DPCO
are f ˆ c ˆ 30 ^ 28 'AM1) [68] and 52.48 'HF/3-
21G) [70]. All the estimated conformations have C2
symmetry.
The dihedral angles of diphenyl ether structures
reported for single crystals [4-19] are plotted in Fig. 1,

K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
409
0
Fig. 3. Shielding constant surfaces for the carbons C4, C6 and C4 of diphenyl ether 'DPO) on the dihedral angles f and c. The contour plot of
the difference between C4 and C6 'D4,6) is also shown at the bottom.
0
and those of benzophenone structures [20±65]2 are
plotted in Fig. 2. These data were collected from the
Cambridge Structural Database [3]. Since the benzene
rings in DPO and DPCO have symmetrical structures,
the conformation of 'c, f) is in same principle as that
of 'f, c). Hence, the conformations of single crystals
are plotted in the area of 08 # f # 908 and c $ f:
The dihedral angles are located around the minimum
region of the surfaces for DPO and DPCO. In addi-
tion, the dihedral angles tend to distribute on the line
of f ˆ c 'Case A) and f 1 c ˆ 908 'Case B) for
DPO, while they are more concentrated on the
minimum region for DPCO.
and C4 ; two dihedral angles have a same value ꢀf ˆ
c†: In contrast, when two split peaks are observed, two
dihedral angles are different ꢀf ± c†:
Fig. 4a shows the cross section of Fig.3 along the
line of f ˆ c: The shielding constant difference
between C4 and C6 'D4,6) varies with the dihedral
angle, which indicates that dihedral angles can be
estimated from the curve of D4,6. Fig. 4b shows the
cross section of Fig. 3 along the line of f ˆ c ˆ 908:
0
0
The shielding constants of C4, C6, C4 ; and C6 ; have
different values in the range0 of 08 , f , 908: In the
0
case of f ˆ 08; C4 and C6 ; have a same shielding
constant, but those of C4 and C6 are different. Clayden
has reported the CP/MAS spectrum of 1,4-diphenoxy-
benzene whose conformation was determined as
ꢀf; c† ˆ ꢀ08; 878† by X-ray diffraction [76]. Two
distinct peaks were observed for C4 and C6,0 while
3.2. Shielding constant calculation
Fig. 3 shows the shielding constant surfaces calcu-
0
0
only one peak was observed for C4 and C6 : This
lated for C4, C6 and C4 of DPO 'see Scheme 2). The
contour plot of the shielding constant difference
between C4 and C6 is also shown at the bottom. The
observation is in good agreement with our shielding
constant calculation.
0
The B3LYP/6-3111G'2d,p) level of theory with
an extended basis set was also used to calculate the
shielding constants of DPO. This level of theory was
highly recommended for nuclear shielding calculation
because the RMS error for 13C chemical shifts relative
to TMS determined using this method is nearly three
times smaller than that determined using RHF/6-
31G'd) [74]. Fig. 5 shows the dihedral angle depen-
dence of the shielding constants of C4, C6 and their
difference 'D4,6) for DPO in Case A ꢀf ˆ c†. The
geometries optimized using B3LYP/6-31G'd) were
shielding constants of C4, C4 and C6 vary with their
dihedral angles as seen in the ®gure. The surfaces of
0
C4 and C4 intersect each other on the line of f ˆ c
'Case A), which means that the chemical shifts of
symmetrical two carbons contained in different
phenyl rings become identical in Case A. As a conse-
quence, conformation of diphenyl region can be
inferred from the number of peaks in NMR spectrum.
That is to say, when only one peak is observed for C4
2
For benzophenone 'DPCO) compounds, see Refs. [20±65]

410
K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
Fig. 5. Comparison of the shielding constants of the carbons C4, C6,
and D4,6 in DPO for f ˆ c calculated using RHF/6-31G'd) 'open
symbols) and B3LYP/6-3111G'2d,p) '®lled symbols) levels of
theory.
Figs. 6a and b show the dihedral angle dependence
of the shielding constants of C4 and C6 of DPCO for
Case A and Case B, respectively. The shielding
constants and their difference 'D4,6) vary with the
dihedral angles, f and c. Hence, the curve of D4,6
can be used to estimate dihedral angles in Case A.
The nuclear shieldings of all carbons show dihedral
angle dependence, but we selected the chemical shifts
of C4 and C6 for estimating the dihedral angles from
the following two reasons. First, the conformation
dependence of the nuclear shieldings of these carbons
is relatively large. Second, the substituent shielding
effects of phthalimide group on these carbons are
negligibly small because these carbons are located
at the meta-position from the group.
Fig. 4. 'a) Cross sections of the surfaces of Fig. 3 along the line of
f ˆ c 'Case A) ꢀ08 # f # 908†; and 'b) those of Fig. 3 along the
line of f 1 c ˆ 908 'Case B). The shielding constant difference
between C4 's4) and C6 's6) is expressed as D4,6
.
used for the calculation. The open and ®lled symbols
show the values obtained from RHF/6-31G'd) and
B3LYP/6-3111G'2d,p), respectively. As seen in the
®gure, two curves obtained from different levels of
theory are similar to each other. The difference
between two calculated values for C6 is less than
0.5 ppm. The largest difference of 1.2 ppm is
observed for C4 at f ˆ 08; and this difference
decreases as f increases. This indicates that the
RHF/6-31G'd) level of theory has suf®cient reliability
for calculating the shielding constants of DPO. Hence,
the RHF/6-31G'd) can be used for the shielding
constant calculations of diphenyl structures.
3.3. Estimation of dihedral angle v
The dihedral angle v, which is de®ned as the
interring rotation of imide ring and phenylene ring,
can be estimated using the correlation between the
differences of solid-state and solution-state chemical
shifts 'Dsolid±solution) of C1 carbon and the dihedral
angles of N-phenylphthalimide derivatives deter-
mined by X-ray crystallography 'Fig. 7) [77]. This
correlation shows that the NMR chemical shift of C1
is displaced to a higher frequency as the v increases

K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
411
Fig. 7. Dependence of the chemical shifts difference 'Dsolid±solution) of
C1 carbon on the dihedral angle v for N-phenylphthalimide deri-
vatives 'Ð: calculated,X: experimental).
Fig. 6. Dihedral angles dependence of the shielding constants of
benzophenone 'DPCO) for the cases of 'a) f ˆ c 'Case A) ꢀ08 #
f # 908†; and 'b) f 1 c ˆ 908 'Case B). The shielding constant
difference between C4 's4) and C6 's6) is expressed as D4,6 in 'a).
from 0 to 908. Dihedral angle v can be estimated by
applying a value of Dsolid±solution for this curve.
Figs. 8±12 show 13C solution-state NMR and solid-
state CP/MAS NMR spectra of BHPI-DPO, BHPI-
DPCO, BPI-DPO, BPI-DPCO and ODPA/ODA,
respectively. The peaks in the spectra were assigned
using substituent shielding effects [78]. The signal
assignment of BPI-DPO reported by Varner et al.
[79], and the peak positions estimated from the
shielding constant calculation were also used to assign
the signals in the diphenyl region.
Fig. 8. 'a) solution, 'b) TOSS, 'c) TOSS & DD and 'd) subtracted 'c
from b) NMR spectra and signal assignments of BHPI-DPO.
The symbols in the spectra correspond to the carbons shown in
Scheme 1.

412
K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
Fig. 10. 'a) solution, 'b) TOSS, 'c) TOSS & DD and 'd) subtracted
'c from b) NMR spectra and signal assignments of BPI-DPO.
Fig. 9. 'a) solution, 'b) TOSS, 'c) TOSS & DD and 'd) subtracted 'c
from b) NMR spectra and signal assignments of BHPI-DPCO.
The solid- and solution-state chemical shifts of C1
is 127.1 and 129.0 ppm, respectively for BHPI-DPO
'Figs. 8a and b). The observed Dsolid±solution value of
21.9 ppm corresponds to v ˆ 54 ^ 58: In the same
way, the dihedral angles of v were estimated for all
the compounds. Table 1 summarizes the estimated
values of v. The NMR spectra of powder and ®lm
samples are completely identical for ODPA/ODA
'Fig. 12). The chemical shift of C1 in solids is
126.7 ppm. The polyimide ODPA/ODA is not soluble
to solvents. However, since BHPI-DPO can be
regarded as the repeating unit of ODPA/ODA, the
solution chemical shift of BHPI-DPO '129.0 ppm)
can be used as the chemical shift of ODPA/ODA in
solution. The v value thus estimated using the
resultant value of Dsolid±solution '22.3 ppm) is 45 ^ 58.
The estimated v values of the samples, in this study
are larger than 458. As seen in Fig. 7, the v values of
NPPI derivatives range from 53 to 908, and no
samples having planar conformations have been
reported. Hence, it is considered that the chemical
bond connecting the imide ring and phenyl ring in
imide compounds always twist signi®cantly. Mean-
while, the values of v may contain errors of ^58
because the peaks in the CP/MAS spectra are broa-
dened compared to the solution spectra, and some
peaks are overlapped. The experimental errors of
chemical shifts are ca. ^0.5 ppm in the usual case.
Table 1
13C Chemical shifts of C1 carbons in the solid and solution-state and
v values estimated from the chemical shift differences 'D)
Sample
Solid 'ppm) Solution 'ppm) D 'ppm) v '8)
BHPI-DPO
127.1
129.0
126.6
127.5
125.8
129.0
2 1.9
2 0.8
3.3
54 ^ 5
58 ^ 5
90 ^ 5
55 ^ 5
45 ^ 5
BHPI-DPCO 125.8
BPI-DPO
BPI-DPCO
131.3
124.8
3.4. Estimation of dihedral angles f and c
2 1.0
2 2.3
ODPA/ODA 126.7
Since single peaks are observed for C2 and C5 in the

K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
413
Fig. 11. 'a) solution, 'b) TOSS, 'c) TOSS & DD and 'd) subtracted
'c from b) NMR spectra and signal assignments of BPI-DPCO.
Fig. 12. NMR spectra and signal assignments of ODPA/ODA. 'a)
TOSS spectrum of powder sample, 'b) TOSS & DD spectrum of
powder sample, 'c) TOSS spectrum of ®lm sample, and 'd) TOSS &
DD spectrum of powder sample.
TOSS & DD spectrum of BHPI-DPO 'Fig. 8c), two
dihedral angles, f and c, are estimated to be iden-
tical. From the subtracted spectrum 'Fig. 8d), the
chemical shift difference 'D4,6) of C4 '122.2 ppm)
and C6 '119.0 ppm) is estimated as 3.2 ppm. By
applying the value of D4,6 to Fig. 4a, the dihedral
angles f and c are estimated as f ˆ c ˆ 39 ^ 58:
Some experimental errors could be included in the
estimation procedure proposed in this study from the
following reasons. First, single molecules of model
compounds were used for the calculation of shielding
constants, but substituent and molecular packing
effects may cause displacements in the observed
solid chemical shifts. Second, errors may occur in
applying experimental chemical shifts to calculated
curves of shielding constants. A small difference in
chemical shift sometimes causes signi®cant difference
in dihedral angle in gently slope regions. The esti-
mated values of dihedral angles generally contain
errors of ^58.
as f ˆ c because a single peak is observed for C2 in
the TOSS&DD spectrum 'Fig. 9c). The value of D4,6
was estimated as 2.1 ppm from the chemical shifts of
C4 'a shoulder at 131.4 ppm) and C6 '129.3 ppm) 'Fig.
9d). The dihedral angles were estimated as f ˆ c ˆ
308 by applying the value of D4,6 '2.1 ppm) to Fig. 6a.
Since single peaks are observed for C2 and C5 in the
TOSS & DD spectrum of BPI-DPCO 'Fig. 11), two
dihedral angles, f and c, are estimated to be iden-
tical. The chemical shift difference 'D4,6) is estimated
as less than 1 ppm because of the overlap of these
signals. Hence, the dihedral angle 'f, c) of BPI-
DPCO are estimated to be larger than 308 by applying
the value of D4,6 to Fig. 6a. On the other hand, the MO
calculations exhibit that the value of D2,5 'difference
between the shielding constants between C2 and C5)
signi®cantly increases as the dihedral angle f
increases. The value of D2,5 is 0.5 ppm for f , 208;
2.0 ppm at 308, 12.3 ppm at 608, and 16.5 ppm at 908.
The conformation of BHPI-DPCO is also estimated

414
K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
Table 2
13C Chemical shift differences between carbons C4 and C6 'D4,6) and
estimated values of dihedral angles f and c
The ratio of hydrogen bonded carbonyl to non
hydrogen bonded carbonyl is estimated as 1 to 1 for
BHPI-DPO from the peak intensities of two carbonyl
peaks at 170 and 165 ppm 'Fig. 8b). This indicates that
all the hydroxyl groups form intermolecular hydrogen
bonds. In contrast, the ratio of the intensity of the
hydrogen bonded central carbonyl '194.8 ppm) to the
non-hydrogen bonded central carbonyl '190.1 ppm) is 4
to 1 for BHPI-BPCO, and the ratio for the imide carbo-
nyls '169.0 and 166.4 ppm) is 1±2 'Fig. 9b). Consid-
ering the numbers of carbonyl and hydroxyl groups, all
the molecules should not form a same type of hydrogen
bonds. It seems that at least two crystal forms exist in
BHPI-DPCO from the intensities of carbonyl peaks.
Two shoulder peaks are observed for C5 in Fig. 9c.
The fact that the numbers of the peaks assigned to the
quaternary carbons is larger than that of the quaternary
carbons also suggests the existence of plural crystalline
forms, but the molecular conformations in the different
forms should be identical within the experimental error
of the procedure proposed in this study.
Sample
D4,6 'ppm)
'f, c) '8)
BHPI-DPO
BHPI-DPCO
BPI-DPO
3.2
'39 ^ 5, 39 ^ 5)
'30 ^ 5, 30 ^ 5)
60±90
2.1
0±1
0±1
0±2
a
BPI-DPCO
ODPA/ODA
'30 ^ 5, 30 ^ 5)
30±90
a
The conformation was estimated not from the value of D4,6 but
from the value of D2,5
.
The experimental value of D2,5 is estimated as
2.1 ppm from the chemical shifts of C2 '133.6 ppm)
and C5 '131.5 ppm). As a consequence, the dihedral
angles are estimated as f ˆ c ˆ 30 ^ 58:
Two dihedral angles are estimated to be identical
ꢀf ˆ c† for BPI-DPO 'Fig. 10) and ODPA/ODA 'Fig.
12) from the spectral shapes, but the dihedral angles
were predicted with large errors as f ˆ c ˆ 60±908
for BPI-DPO and f ˆ c ˆ 30±908 for ODPA/ODA
because the exact chemical shifts of C4 and C6 cannot
be obtained due to the spectral overlap. In particular,
the broadened signals of C2 and C5 in Fig.12 suggest a
large distribution of f and c in OPDA/ODA polyi-
mide.
4. Conclusions
Solution- and solid-state 13C NMR have been used
to investigate the conformations of bis'4-hydro-
xyphthalimde)s. The dihedral angles f, c, and v
were estimated using the spectral shapes 'number of
peaks) and the observed chemical shifts. From the
shielding constant calculations, it was clari®ed that
the number of peaks of the carbons in diphenyl region
can be used to estimate the relationship between two
dihedral angles f and c. When the signal of each
carbon is singlet, two dihedral angles f and c are
identical, otherwise they are different. Moreover, the
values of f and c were estimated by comparing the
13C NMR chemical shift differences of phenyl carbons
with the variations in the calculated shielding
constants. The comparison of nuclear shielding calcu-
lations between RHF/6-31G'd) and B3LYP/6-
3111G'2d,p) levels of theory indicates that RHF/6-
31G'd) has suf®cient reliability for calculating
shielding constants of diphenyl structures. It was
revealed that the dihedral angle v ranges between
45 and 908, and the dihedral angles 'f, c) distributes
on the stable region of the energy surface ranging
from 30 to 908. All the molecules studied have
Table 2 summarizes the estimated values of f and
c. The estimated conformation of f ˆ c ˆ 39 ^ 58
for BHPI-DPO is located in the most stable region in
the conformational energy surfaces of DPO 'Fig. 1).
In addition, the conformation of f ˆ c ˆ 30 ^ 58 for
BHPI-DPCO is also located in the most stable region
of DPCO 'Fig. 2).
3.5. Peak splitting of imide carbonyl carbons
A splitting of ca. 3 ppm is observed for the carbonyl
signal in the TOSS spectra of BHPIs. On the other
hand, no such splittings are observed in the TOSS
spectra of BPIs. Since BHPIs have two hydroxyl
groups at the termini of molecule and have four or
®ve carbonyl groups, these splittings should be caused
by intermolecular hydrogen bonds. The peak splitting
observed for the benzophenone carbonyl carbon of
BHPI-DPCO indicates that hydrogen bonds are
formed not only with the imide carbonyl but also
with the central carbonyl group.

K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
415
[19] T. Nevalainen, K. Rissanen, J. Chem. Soc., Perkin Trans. 2
'1994) 271 [YEBGOB].
symmetrical conformations. The carbonyl peaks
displaces to higher frequency by 3±5 ppm by forming
intermolecular hydrogen bond. The ratio of hydrogen
bonded imide carbonyl carbons to non-hydrogen
bonded imide carbonyls is 1 to 1 for BHPI-DPO and
1±2 for BHPI-DPCO.
[20] J.D. Miscovic, S.M. Ramaseshan, Glasn. Khem. Drushtva
Beograd 'Bull. Soc. Chim. Beo.) 32 '1967) 253
[AMBBPO].
[21] Z. Skrzat, A. Rozak, Acta Crystallogr., Sect. B 37 '1981) 770
[AMBZAC].
[22] G.P.M. van der Velden, J.H. Noordik, J. Cryst. Mol. Struct. 10
'1980) 83 [AMBZPH].
[23] J.D. Miscovic, S.M. Ramaseshan, Glasn. Khem. Drushtva
Beograd 'Bull. Soc. Chim. Beo.) 32 '1967) 253 [AMCBPO].
[24] E.O. Schlemper, Acta Crystallogr., Sect. B 38 '1982) 554
[BADVIL10].
Acknowledgements
We wish to thank S. Sasaki and T. Matsuura at NTT
for providing 4-hydroxyphthalimide.
[25] T. Ray, S.P. Sen Gupta, Cryst. Struct. Commun. 10 '1981)
1123 [BAGPAA].
[26] J. Delettre, C. R. Acad. Sci., Ser. C 269 '1969) 103
[BOBZAC].
References
[27] R.A. Lalancette, P.A. Vanderhoff, H.W. Thompson, Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 46 '1990)
1682 [BOBZACO2].
[1] H. Lee, D. Stoffery, K. Neville, New Linear Polymers,
McGraw-Hill, New York, 1967.
[28] G.M. Lobanova, Kristallogra®ya 13 '1968) 984
[BPHENO10].
[2] S. Sasaki, Y. Hasuda, J. Polym. Sci. Polym. Lett. 25 '1987)
377.
[29] E.B. Fleischer, N. Sung, S. Hawkinson, J. Phys. Chem. 72
'1968) 4311 [BPHENO10].
[3] F.H. Allen, O. Kennard, Chem. Des. Automn. News 8 '1993)
31.
[30] V. Pattabhi, K. Venkatesan, J. Cryst. Mol. Struct. 3 '1973) 25
[BRBEPH10].
[4] R. Capon, E.L. Ghisalberti, P.R. Jefferies, B.W. Skelton, A.H.
1 '1981) 2464
White, J. Chem. Soc., Perkin Trans.
[BAJVOX].
[31] Z. Skrzat, A. Konitz, Pol. J. Chem. 54 '1980) 1029
[BROHBZ].
[5] E.A.H. Grif®th, W.D. Chandler, B.E. Robertson, Can. J.
Chem. 50 '1972) 2979 [CMNPCB].
[32] K.G. Shields, C.H.L. Kennard, J. Chem. Soc., Perkin Trans. 2
'1977) 463 [CBENPH].
[6] J. Eriksson, L. Eriksson, E. Jakobsson, Acta Crystallogr., Sect.
C: Cryst. Struct. Commun. 55 '1999) 2169 [COCDED].
[7] P. Singh, J.D. McKinney, Acta Crystallogr., Sect. B 36 '1980)
210 [DCPHER].
[33] M.M. Granger, M.F. Coillot, Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 41 '1985) 542 [CBENPH01].
[34] Z. Skrzat, Acta Crystallogr., Sect.
[CLOHBZ].
B 36 '1980) 2812
[8] E.A.H. Grif®th, W.D. Chandler, B.E. Robertson, Can. J.
Chem. 50 '1972) 2972 [DNPBUB].
[35] B. Peric, B. Kojic-Prodic, Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 56 '2000) 211 [CUZKUD01].
[9] C.H.L. Kennard, G. Smith, T. Hari, Aust. J. Chem. 40 '1987)
1131 [FIPXIL].
[36] L. Antolini, I.M. Vezzosi, L.P. Battaglia, A.B. Corradi,
J. Chem. Soc., Perkin Trans. 2 '1985) 237 [DEMBAY].
[37] B.W. Liebich, Acta Crystallogr., Sect. B 35 '1979) 1186
[DHXBZP10].
[10] K. Prout, J. Fail, R.M. Jones, R.E. Warner, J.C. Emmett,
J. Chem. Soc., Perkin Trans. 2 '1988) 265 [GAPWUP].
[11] K. Rissanen, J. Valkonen, L. Virkki, Acta Crystallogr., Sect.
C: Cryst. Struct. Commun. 44 '1988) 1644 [GICLUZ].
[12] K. Rissanen, L. Virkki, Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 45 '1989) 1408 [KARJES].
[38] H.G. Norment, I.L. Karle, Acta Crystallogr. 15 '1962) 873
[DMBOPN].
[39] O.A. Dvorkin, S.A. Andronati, T.Sh. Gifeisman, Yu.O.
Simonov, O.S. Yavors'kii, V.I. Pavlov'skii, Dopov. Akad
Nauk Ukr. RSR Ser. B-Geol. Khim. Biol. Nauki Aug-34
'1985) [DUDZIL].
[13] E.A.H. Grif®th, W.D. Chandler, B.E. Robertson, Can. J.
Chem. 50 '1972) 2963 [MNPPRB].
[14] R. Shankar, P. Sharma, A. Cabrera, G. Espinosa, N. Rosas,
N.K. Jha, P. Vasudevan, Indian J. Chem., Sect. B 35 '1996)
894 [NUVPID].
[40] P.J. Cox, A.T.Md. Anisuzzaman, R.H. Pryce-Jones, G.G.
Skellern, A.J. Florence, N. Shankland, Acta Crystallogr.,
Sect. C: Cryst. Struct. Commun. 54 '1998) 856 [DUDZIL01].
[41] Y. Ito, T. Matsuura, K. Tabata, M. Ji-Ben, K. Fukuyama, M.
Sasaki, S. Okada, Tetrahedron 43 '1987) 1307 [FEVMUO].
[42] H. Kutzke, M. Al-Mansour, H. Klapper, J. Mol. Struct. 374
'1996) 129 [FEVNAV02].
[15] B. Krumm, A. Vij, R.L. Kirchmeier, J.M. Shreeve, Inorg.
Chem. 36 '1997) 366 [RIBTAX].
[16] L.A. Chetkina, A.N. Sobolev, V.K. Bel'skii, L.E. Serdyuk,
I.E. Kardash, Kristallogra®ya 36 '1991) 902 [SUCVER].
[17] U. Orn, L. Eriksson, E. Jakobsson, A. Bergman, Acta Chem.
Scand. 50 '1996) 802 [TOQBAC].
[43] O.A. Dvorkin, T.Sh. Gifeisman, S.A. Andronati, V.I.
Pavlovs'kii, O.S. Yavors'kii, T.I. Malinovs'kii, Dopov.
Akad. Nauk Ukr. RSR Ser. B-Geol. Khim. Biol. Nauki
'1986) 36 [FINCEK].
[18] J.D. McKinney, S.C. Agarwal, W.T. Padgett, G.R. Lambert,
P. Singh, S. Nesnow, Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 51 '1995) 428 [WINLOU].

416
K. Aimi et al. / Journal of Molecular Structure 602±603 /2002) 405±416
[44] S.B. Martin, E.J. Valente, J. Chem. Cryst. 28 '1998) 203
[GEQTUR].
[63] S. Gen-Bo, P. Feng, Li. Zheng-Dong, He. You-Ping, H. Liang-
Ren, J. Ri-Hong, Chin. J. Struct. Chem. 'Jiegou Huaxue) 10
'1991) 171 [VOFVAN].
[45] S.J. Maginn, R.J. Davey, Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 50 '1994) 254 [HATXIJ].
[64] S. Guo, G. Su, F. Pan, Y. He, Z. Kristallogr. 202 '1992) 296
[WADBEI].
[46] B. Berger, M. Bolte, Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 50 '1994) 773 [HECHEC].
[65] G.P.M. van der Velden, J.H. Noordik, J. Cryst. Mol. Struct. 9
'1979) 283 [ZZZOVY01].
[47] D. Matak, M. Vinkovic, M. Dumic, Acta Crystallogr., Sect. C:
Cryst. Struct. Commun. 50 '1994) 1339 [HEKYOL].
[48] S.I. Popov, T.M. Kopylova, A.M. Andrievsky, S.V.
Medvedev, A.V. Yatsenko, Zh. Org. Khim. 30 '1994) 266
[HEXDIX].
[66] J. Kendrick, J. Chem. Soc., Faraday Trans. 86 '1990) 3995.
[67] A. Anwer, R. Lovell, A.H. Windle, in: R.J. Roe 'Ed.),
Computer Simulation of Polymers, Prentice Hall, New Jersey,
1991 chapter 3.
[49] B.W. Liebich, Acta Crystallogr., Sect. B 32 '1976) 431
[HMXCBP10].
[68] M.P. Makowski, Comput. Polym. Sci. 3 '1993) 1.
[69] T. Straûner, Can. J. Chem. 75 '1997) 1011.
[50] A. Degen, M. Bolte, Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 55 '1999) 1306 [HOFGAK].
[70] M. Feigel, J. Mol. Struct. 'THEOCHEM) 366 '1996) 83.
[71] A.E. Tonelli, D.B. Chesnut, Macromoelcules 29 '1996) 7.
[72] S. Ando, T. Hironaka, H. Kurosu, I. Ando, Magn. Reson.
Chem. 38 '2000) 241.
[51] P. Briard, J.C. Rossi, Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 46 '1990) 1036 [KEMRUP].
[52] H. Preut, T. Kolev, I. Juchnovski, P. Bleckmann, H, Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 48 '1992) 938
[KOTBEA].
[73] R. Ditch®eld, Mol. Phys. 27 '1974) 789.
[74] J.R. Cheeseman, G.W. Trucks, T.A. Keith, M.J. Frisch,
J. Chem. Phys. 104 '1996) 14.
[53] L. Wenjie, S. Genbo, H. Yougping, Acta Crystallogr., Sect. C:
Cryst. Struct. Commun. 50 '1994) 1638 [LEZKIK].
[54] G. Chiari, H.C.R. Taylor, F.R. Fronczek, G.R. Newkome, G,
Acta Crystallogr., Sect. B 36 '1980) 2488 [NOPHKN].
[55] P.J. Cox, A.T.Md. Anisuzzaman, R.H. Pryce-Jones, G.G.
Skellern, A.G. Florence, N. Shankland, Acta Crystallogr.,
Sect. C: Cryst. Struct. Commun. 54 '1998) 856 [NUBDAP].
[56] N.L. Vasco-Mendez, K. Panneerselvam, E. Rudino-Pinera, M.
Soriano-Garcia, Anal. Sci. 12 '1996) 677 [NUVFAL].
[57] S. Ebbinghaus, D. Abeln, M. Epple, Z. Kristallogr. New Crys.
Struct. 212 '1997) 339 [PIQLIK].
[75] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A.
Robb, J.R. Cheeseman, V.G. Zakrzewski, J.A. Montgomery,
R.E. Stratmann, Jr., J.C. Burant, S. Dapprich, J.M. Millam,
A.D. Daniels, K.N. Kudin, M.C.C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A. Petersson,
P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D.
Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski,
J.V. Ortiz, A.G. Baboul, B.B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox,
T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C.
Gonzalez, M. Challacombe, P.M.W. Gill, B. Johnson, W.
Chen, M.W. Wong, J.L. Andres, C. Gonzalez, M. Head-
Gordon, E.S. Replogle, J.A. Pople, gaussian 98, Revision
A.9, Gaussian, Inc., Pittsburgh, P.A., 1998.
[58] Z. Skrzat, Pol. J. Chem. 54 '1980) 795 [QQQHDS10].
[59] Su Genbo, P. Feng, He Youping, Acta Crystallogr., Sect. C:
Cryst. Struct. Commun. 48 '1992) 210 [TANSEG].
[60] G. Shouwu, Su Genbo, P. Feng, He Youping, Acta Crystal-
logr., Sect. C: Cryst. Struct. Commun. 48 '1992) 576
[TAPVAH].
[76] N.J. Clayden, J. Chem. Soc., Perkin Trans. 2 '1990) 729.
[77] H. Ishii, I. Ando, S. Ando, NMR prep. Jpn. 37 '1998) 332.
[78] D.F. Ewing, Org. Magn. Reson. 12 '1979) 499.
[79] S.J. Varner, R.L. Vold, G.L. Hoatson, Solid State Nucl. Magn.
Reson. 12 '1998) 71.
[61] [TAPVAH01] S. Ebbinghaus, D. Abeln, M. Epple, private
communication, '1999).
[62] P.J. Cox, A.T.Md. Anisuzzaman, G.G. Skellern, R.H. Pryce-
Jones, A.J. Florence, N. Shankland, Acta Crystallogr., Sect. C:
Cryst. Struct. Commun. 53 '1997) 476 [TUDNUB].