2278 Watanabe et al.
Macromolecules, Vol. 35, No. 6, 2002
HCl‚TEA salt precipitated immediately. The reaction was kept
for another 2 h at 5 °C and then 24 h at 60 °C. The reaction
mixture was filtered and concentrated at reduced pressure.
The crude product was distilled under reduced pressure to give
2a as a clear oil. The yield was 10.1 g (52%): bp 88-94 °C
(0.75 Torr). IR (NaCl): ν ) 3409 (N-H), 1249 (Si-CH3), 836
(Si-CH3) cm-1. Anal. Calcd for (C16H38N2Si2): C, 61.07; H,
12.17; N, 8.90. Found: C, 60.99; H, 12.32; N, 9.12.
2,5(2,6)-Bis(N-tr im eth ylsilylam in om eth yl)bicyclo[2.2.1]-
h ep ta n e (2b). This compound was prepared as described
above using 2,5(2,6)-bis(aminomethyl)- bicyclo[2.2.1]heptane
in place of 5-amino-1,3,3-trimethylcyclohexanemethylamine.
The yield of 2b was 48% as a clear oil: bp 106-110 °C (0.75
Torr). IR (NaCl): ν ) 3405 (N-H), 1249 (Si-CH3), 836 (Si-
CH3) cm-1. Anal. Calcd for (C15H34N2Si2): C, 60.33; H, 11.48;
N, 9.38. Found: C, 59.95; H, 11.14; N, 9.66.
5-ter t-Bu tyld im eth ylsilyla m in o-N-ter t-bu tyld im eth yl-
silyl-1,3,3-tr im eth ylcycloh exa n em eth yla m in e (2c). This
compound was prepared as described above using tert-butyl-
dimethylchlorosilane in place of trimethylchlorosilane. The yield
of 2c was 82% as a clear oil: bp 120-123 °C (0.15 Torr). IR
(NaCl): ν ) 3405 (N-H), 1253 (Si-tert-Butyl), 829 (Si-tert-
Butyl) cm-1. Anal. Calcd for (C22H50N2Si2): C, 66.26; H, 12.64;
N, 7.02. Found: C, 66.07; H, 12.32; N, 7.26.
4,4′-Met h ylen eb is(N,N′-ter t-b u t yld im et h ylsilylcyclo-
h exyla m in e) (2d ). This compound was prepared as described
above using 4,4′-methylenebis(cyclohexylamine) in place of
5-amino-1,3,3-trimethylcyclohexanemethylamine. The yield of
2d was 50% as a clear oil: bp 144 °C (0.02 Torr). IR (NaCl):
ν ) 3397 (N-H), 1249 (Si-tert-Butyl), 829 (Si-tert-Butyl)
cm-1. Anal. Calcd for (C25H54N2Si2): C, 68.42; H, 12.40; N, 6.38.
Found: C, 68.13; H, 12.11; N, 6.37.
(d,p)), which is implemented in the “Gaussian-98” (Revision
A.7) program, was used to calculate molecular orbital energies
of dianhydrides. The geometric structures of computed com-
pounds were fully optimized under no constraints prior to the
calculations of molecular orbital energies.
Resu lts a n d Discu ssion
Syn th esis of P oly(a m ic a cid tr ia lk ylsilyl ester s)
(3). As alicyclic dianhydrides (1), 1,2,3,4-cyclobutane-
tetracarboxylic dianhydride (1a ) and bicyclo[2.2.1]-
heptane-2-methanecarboxylic-3,5,6-tricarboxylic-2,3:5,6-
dianhydride (1b) were employed. On the other hand,
N-trimethylsilylated alicyclic diamines (2), 5-trimethyl-
silylamino-N-trimethylsilyl-1,3,3-trimethylcyclohexane-
methylamine (2a) and 2,5(2,6)-bis(N-trimethylsilylamino-
methyl)bicyclo[2.2.1]heptane (2b) were prepared by the
reactions of corresponding diamines with trimethyl-
chlorosilane in the presence of TEA.
To determine optimal conditions for polymerization,
the polymerization of 1a with N-silylated diamines 2a
or 2b was studied in detail. The ring-opening poly-
addition was performed with 1 mmol of monomers in a
solvent at room temperature for 1 h [eqs 1 and 2]
P oly(a m ic a cid tr ia lk ylsilylester )s (3). A flame-dried 25
mL flask was charged with 2a (0.899 g, 2.86 mmol) and DMAc
(3.41 g) under nitrogen. 1,2,3,4-Cyclobutanetetracarboxylic
dianhydride 1a (0.560 g, 2.86 mmol) was added to this solution
at 25 °C in one portion. The solution was stirred at this
temperature for 1 h under nitrogen. The inherent viscosity of
the resulting polymer solution was 0.99 dL/g, measured at a
concentration of 0.5 g/dL in DMAc at 30 °C. IR (KBr): ν ) 1708
(CdO, silylester), 1654 (CdO, amide) cm-1
.
P oly(a m ic a cid )s (4). The solution of polymer 3 was poured
into a 0.2 wt % aqueous HCl solution. The precipitate was
filtered off, washed with water, and dried under vacuum.
IR (KBr): ν) 1727 (CdO, carboxylic acid), 1646 (CdO,
amide) cm-1. Anal. Calcd for (C18H26N2O6‚0.8H2O)n: C, 56.77;
H, 6.88; N, 7.36. Found: C, 56.76; H, 6.51; N, 7.36.
P oly(im id e)s (5). The polyimide film was prepared by
casting the solution of polymer 3 on a silicon wafer at room
temperature and then by heating at 100, 150, 200, and 250
°C for 0.5 h and 300 °C for 1 h under nitrogen atmosphere.
IR (KBr): ν ) 1770 and 1708 (imide CdO), 1361 (C-N) cm-1
.
Anal. Calcd for (C18H22N2O4‚0.1H2O)n: C, 65.08; H, 6.74; N,
8.43. Found: C, 64.54; H, 6.58; N, 8.40.
Mea su r em en ts. The infrared spectra were recorded on a
Horiba FT-720 spectrophotometer. 1H NMR spectra were
obtained using a Bruker DPX300 (300 MHz) spectrometer. The
UV-visible spectra of polyimide films were recorded on a J asco
V-560 spectrophotometer. Thermogravimetry (TG) and dif-
ferential scanning calorimetry (DSC) were performed with a
Seiko TG/DTA 6300 and DSC 6200, respectively. Molecular
weights were determined by a gel permeation chromatograph
(GPC) with polystyrene calibration using TOSO HPLC (HLC-
8120GPC) equipped with a TOSO TSK gel column (TSKgel
GMHHR-M, TSKgel GMHHR-L) at 40 °C in DMF with LiBr.
Refractive indices of polyimide films formed on quartz sub-
strates were measured at a wavelength of 1.320 µm at room
temperature with a Metricon model PC-2000 prism coupler.
Using linearly-polarized laser light with parallel (TE: trans-
verse electric) and perpendicular (TM: transverse magnetic)
polarization to the film plane, the in-plane (nTE) and out-of-
plane (nTM) refractive indices and the film thickness of the
samples were determined.
Table 1 summarizes the effects of solvent on the
inherent viscosity of polymer 3. Polymerizations of 2a
or 2b with 1a proceeded rapidly in a homogeneous
solution without salt formation and were completed in
1 h, giving colorless and viscous polymer solutions. The
polymers with inherent viscosities ranging between 0.71
and 0.99 dL/g were obtained in DMAc.
On the basis of these results, the polymerization of
2a or 2b with 1b was performed in DMAc at room
temperature. Polymers 3ba or 3bb obtained after 24 h
showed very low inherent viscosities of 0.09 and 0.12
dL/g. The polymerization was carried out at 60 °C again,
but no discernible effect on the inherent viscosity was
observed (Table 2). This result would be explained by a
low reactivity of 1b compared to that of 1a . Ando et al.22
have reported that the acylation rate constants of
Ca lcu la tion s. A density functional level of theory (B3LYP)
with an extended basis set with polarization functions (6-31G-