Reevaluation of Tetrahydrophthalic Anhydride as an End Cap for Improved Oxidation Resistance in Addition Polyimides

Article abstract of DOI:10.1021/ma0353078

Several substituted 1,2,3,6-tetrahydrophthalic anhydride end caps, including the 3-phenyl, 3-methoxy, 3-trimethylsilyloxy, and 3,6-diphenyl analogues, were synthesized via the Diels-Alder condensation of the corresponding butadienes and maleic anhydride. These anhydrides, as well as the commercially available 3-hydro and 4-methyl analogues, were each ground up together with methylenedianiline in a 2:1 ratio and heated gradually from 204 to 371°C, with the thermolysis followed by NMR. Generally speaking, a transformation via monoimide to bisimide was observed in the lower temperature range, followed by competition between cross-linking and aromatization. We believe that this competition produces a substantial percentage of aromatic product, with the concomitant lowering of the relative amount of cross-linking and is responsible for both improved thermooxidative stability of tetrahydrophthalic end-capped polyimides and their substantial frangibility. The thermolysis of the tetrahydrophthalimides under an inert atmosphere dramatically lowers the amount of aromatization; hence, the mechaniam for aromatization is an oxidative one.

Full text of DOI:10.1021/ma0353078

Macromolecules 2004, 37, 1289-1296
1289
Reevaluation of Tetrahydrophthalic Anhydride as an End Cap for
Improved Oxidation Resistance in Addition Polyimides
Ma r y An n B. Mea d or ,*,† Ar yeh A. F r im er ,‡ a n d J . Ch r istop h er J oh n ston †
NASA Glenn Research Center, Cleveland, Ohio 44135, and Department of Chemistry,
Bar-Ilan University, Ramat Gan 52900, Israel
Received September 4, 2003
ABSTRACT: Several substituted 1,2,3,6-tetrahydrophthalic anhydride end caps, including the 3-phenyl,
3-methoxy, 3-trimethylsilyloxy, and 3,6-diphenyl analogues, were synthesized via the Diels-Alder
condensation of the corresponding butadienes and maleic anhydride. These anhydrides, as well as the
commercially available 3-hydro and 4-methyl analogues, were each ground up together with methylene-
dianiline in a 2:1 ratio and heated gradually from 204 to 371 °C, with the thermolysis followed by NMR.
Generally speaking, a transformation via monoimide to bisimide was observed in the lower temperature
range, followed by competition between cross-linking and aromatization. We believe that this competition
produces a substantial percentage of aromatic product, with the concomitant lowering of the relative
amount of cross-linking and is responsible for both improved thermooxidative stability of tetrahydro-
phthalic end-capped polyimides and their substantial frangibility. The thermolysis of the tetrahydro-
phthalimides under an inert atmosphere dramatically lowers the amount of aromatization; hence, the
mechanism for aromatization is an oxidative one.
In tr od u ction
Fiber-reinforced high-temperature polyimide matrix
composites offer significant advantages in structural
applications over other materials because of their low
density and high specific strength. These composites are
attractive for use in aerospace systems, e.g., aircraft
engines, airframe, missiles, and rockets, where weight
is critical. This weight reduction has substantial benefits
in terms of fuel savings, increased passenger or cargo
load, or increased speed and maneuverability. The
durability and reliability of materials used in aerospace
components is a critical concern. Among the materials
requirements for these applications are a high glass-
transition temperature, Tg (at least 25 °C higher than
the intended use temperature), good high-temperature
stability in a variety of environments, and good me-
chanical properties over a wide range of temperatures.
In general, the stability and/or Tg of most organic
polymers limit their use, at best, to applications in
which temperatures are not higher than 300 °C.
Addition-curing polyimides were investigated to im-
prove the processability of condensation polyimides
without adversely affecting their stability and high-
temperature performance. The most noteworthy devel-
opment is the polymerization of monomer reactant
(PMR) family of polyimides, in particular PMR-15.1
Figure 1 is the typical reaction scheme for the formation
of these end-capped addition polyimides (PMR-15). For
PMR-15, reinforcement fibers are impregnated with a
solution of the dialkyl ester of 3,3′,4,4′-benzophenone-
tetracarboxylic diacid diester (BTDE), methylenedi-
aniline (MDA), and the monomethyl ester of 5-nor-
bornene-2,3-dicarboxylic acid (nadic acid ester, NE) in
a low boiling solvent, typically methanol or ethanol. In
F igu r e 1. Reaction scheme for the preparation of fully cross-
linked PMR-15.
the first step of this process, these monomers undergo
imidization at temperatures around 200 °C to yield
short-chain norbornene end-capped polyimide oligomers
(formulated molecular weight ) 1500 g/mol for PMR-
15). At temperatures above 300 °C, these oligomers
undergo a cross-linking reaction involving the nor-
bornene end cap.2
Ironically, however, it is this very end cap, so impor-
tant to processing, that accounts for much of the weight
loss in the polymer on aging in air at elevated temper-
atures (>315 °C). Thus, the end cap limits the use of
PMR polyimides to lower temperature parts of the
engine and/or shorter lifetimes. Researchers have tried
to improve the stability by a variety of means, but most
prominently by utilizing structures with more aromatic
† NASA Glenn Research Center.
‡ Ethel and David Resnick Chair in Active Oxygen Chemistry,
Bar-Ilan University.
* Corresponding
author:
Fax
219-977-7132;
e-mail
maryann.meador@nasa.gov.
10.1021/ma0353078 CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/20/2004

1290 Meador et al.
Macromolecules, Vol. 37, No. 4, 2004
gated by TRW and NASA in the 1980s and found to
yield composites with TOS values better than PMR-15,
but which were quite frangible.9 St. Clair and St. Clair10
tried 7 as an end cap for polyimide adhesives. However,
they reported that no cross-linking occurred in this
system until 415 °C, as evidenced by differential scan-
ning calorimetry (DSC). If the cross-linking temperature
could be lowered seither by substitution on the ring or
by catalysissthis might be a viable replacement for the
norbornene end cap. Indeed, onset of decomposition for
polymers made with this was shown to be at a higher
temperature than for the norbornenyl end cap.
F igu r e 2. End-cap degradation pathways (f ) labeled
carbons).
character, e.g., Cycap, V-cap, PEPA, acetylene-termi-
nated imides (ATI), benzocyclobutenes, and biphen-
ylenes.3 However, each modification resulted in changes
in the nature of the cross-linking, undesirable changes
in processability and/or mechanical behavior, lower Tg,
and, at times, sizable increases in cost (e.g., Cycap).
Over the past decade, we have been exploring various
aspects of this thermooxidative degradation in the hope
of gleaning clues on how to design new end caps which
will slow down this degradation and prolong the use
lifetime of polyimide materials. Several years ago, we
reported a number of studies on the thermooxidative
aging of a modification of PMR-15, in which we 13C-
labeled the methyne carbon R to the carbonyl groups.4-6
On the basis of this work, we concluded that this
oxidation proceeds through two primary pathways
(Figure 2).
Path A degradation proceeds through initial opening
of the norbornyl ring to form a biradical which under-
goes attack of oxygen to form a 2-hydroxy-substituted
maleimide 2. Path B degradation proceeds through
oxidation of the bridging methylene of the norbornene
moieties followed by carbon monoxide extrusion. Aro-
matization of the intermediate biradical leads to sub-
stituted quinone 3 or phthalimide 4 and related sec-
ondary degradation products.
Path A oxidation products like 2 are cleavage products
that are most likely formed concomitant with large
amounts of weight loss in the polymer system. In
contrast, structures like 3 and 4 from path B are formed
with very little weight loss. Therefore, new end-cap
structures that more strongly favor path B degradation
should lead to lower weight loss in addition polyimides
and result in less shrinkage and cracking in the oxida-
tion layer.4
Recently, we reported on our design of new end caps
that might favor path B degradation.7,8 We proposed to
utilize structures like 5 where X is more labile than the
C-7 methylene of the nadic end cap or is easily oxidized
to a more labile group under aging conditions. In
particular, we studied the 7-hydroxy analogue 6 (i.e.,
5, X ) CHOH). We provided evidence that, in the air-
aging polymer, the hydroxy-bearing carbon oxidizes to
carbonyl and should, therefore, more highly favor path
B degradation. Indeed, the solid NMR spectral changes
observed are all consistent with rapid conversion of the
aliphatic cross-link to an aromatic one via path B or
bridge degradation.
Exp er im en ta l Section
Solution NMR spectra were obtained on a Varian 400 and
Bruker 200 and 300 MHz Fourier transform spectrometers,
using TMS as the internal standard. Assignments were
facilitated by double resonance, APT, Hetcor, COSY, and
NOESY experiments. The carbon numbering of the various
compounds used in the spectral assignments is shown in eqs
1 and 2. Infrared absorptions were determined with a Nicolet
510P FTIR. High-resolution mass spectra (HRMS) were run
on a VG-Fison AutoSpecE high-resolution spectrometer. The
following TA Instruments machines were used for standard
thermal analyses: Q 1000 Reversible DSC; 2910 DSC; 2940
TMA; 2950 Hi-Res TGA. The TGA/MS/IR system used was
comprised of a TA Instruments 2950 TGA whose effluent gas
was split, one arm leading via a heated (ca. 240 °C) transfer
line to a Nicolet Nexus 470 FTIR, while the other arm led via
a heated (ca. 240 °C) deactivated capillary to a Finnigan
Polaris Q mass spectrometer. Analytical thin-layer chroma-
tography (TLC) was performed using Merck silica gel micro-
cards. Maleic anhydride (Aldrich), cis-1,2,3,6-tetrahydro-
phthalic anhydride (10a , Aldrich), and 4-methyl-cis-1,2,3,6-
tetrahydrophthalic anhydride (10e, TCI-EP) are commercially
available and were used as supplied. Compounds 10b,11 10c,12
and 10f13 are known in the literature. Nevertheless, the 13C
spectral data are unreported, while the 1H NMR data are
missing, inaccurate or of relatively low resolution; hence, these
are cited below where appropriate.
3-P h en yl-cis-1,2,3,6-t et r a h yd r op h t h a lic An h yd r id e
(10b ). The title compound was prepared by refluxing equi-
molar amounts of trans-1-phenyl-1,3-butadiene14 (8b, 10.9 g,
84 mmol) and maleic anhydride (8.2 g, 84 mmol) in toluene
(225 mL) for 16 h under nitrogen. The orange solution was
placed in the freezer overnight to give a yellow-brown precipi-
tate. The latter was filtered and air-dried to give 10b as a light
yellow powder (14.8 g, 65 mmol, 77% yield); mp 119.2-120.8
°C. The mother liquor was evaporated, and the residue was
recrystallized from ether to give an additional 2 g (total yield
87%); mp 116.8-117.8. The product was purified by treating
it with charcoal and recrystallizing from boiling ethyl acetate/
hexane. Colorless product (13 g, 57 mmol; 68% overall yield)
was obtained; mp 121.5-122.5 °C (lit.11 120-121 °C).
10b. 1H NMR (CDCl3): δ 7.37-7.24 (m, 3H, Hp and Hm),
7.21-7.17 (m, 2H, Ho), 6.222 (m, 1H, H4), 6.149 (m, 1H, H5),
3.803 (m, 1H, H3), 3.545 (dd, J ) 10 and 7 Hz, 1H, H2), 3.453
(td, J ) 10 and 2 Hz, 1H, H1), 2.896 (dm, J ) 18 Hz, 1H, H6-
cis to Ph), 2.426 (m, 1H, H6-trans to Ph). 13C NMR (CDCl3): δ
173.766 (C7), 170.699 (C8), 137.321 (Cipso), 129.729 (C4), 128.742
(Cm), 128.659 (Co), 127.907 (C5), 127.596 (Cp), 45.822 (C2),
We now report our studies on 1,2,3,6-tetrahydro-
phthalimide 7 (R ) R′ ) R′′ ) H), an end cap that
contains no X group at all. This structure was investi-

Macromolecules, Vol. 37, No. 4, 2004
Tetrahydrophthalic Anhydrides as End Caps 1291
40.144 (C3), 39.051 (C1) 22.335 (C6). IR (KBr): 1845, 1778, and
1708 (anhydride), 1658 cm-1 (CdC). HRMS (DCI, CH4): calcd
(C14H12O3, M+) 228.0786; obsd 228.0780.
P r ep a r a tion of Bisim id es 12, 15, 17, a n d 21. Anhydride
(10, phthalic anhydride or succinic anhydride 22) and diamine
(16a or b) in a 2:1 molar ratio were ground up together and
heated at 204 °C (400 °F) for 1 h. As determined by NMR, in
nearly each case, the primary product at this stage was the
desired bisimide. However, cyclohexene bisimide 12 was
accompanied by a small amount (generally <2%) of the
intermediate monoimide 11 (see eq 1). (The spectrum of the
latter is quite distinctive because of its AA′XX′ pattern of the
H-10′ and H-11′ aromatic protons at ca. 6.6 and 7.0 ppm,
respectively.) However, in the case of 10b and 10f, the ratio
of monoimide:bisimide (11:12) was 2:1 and 3:1, respectively;
bisimide 12 was the primary product when the reaction
mixture was heated further at 315 °C for 0.5 h. In the case of
10d , the anhydride and MDA were dissolved in THF and the
solvent evaporated; NMR analysis revealed the presence of
bisimide 12d . Heating to 204 °C resulted in a 34% weight loss,
accompanied by the disappearance of the distinctive trimeth-
ylsilyloxy peak in the NMR of the residual solid. The NMR
and HRMS data strongly suggest that the product obtained
is the 3-hydroxybisimide 12g.
3-Met h oxy-cis-1,2,3,6-t et r a h yd r op h t h a lic An h yd r id e
(10c).12 The title compound was prepared by dissolving
equimolar amounts of trans-1-methoxy-1,3-butadiene (8c, Al-
drich, 5.0 g, 59 mmol) and maleic anhydride (5.8 g, 59 mmol)
in acetonitrile (100 mL). The reaction mixture was stirred at
room temperature for 24 h under nitrogen. The solvent was
rotary evaporated and the resulting light yellow liquid (10.6
g; 58 mmol; 98% yield) solidified upon standing. The latter
was gradually dissolved in 10 portions of hot ether (50-75 mL
each), and the resulting solution concentrated down to 150 mL,
at which time crystallization began to give the desired product
(9.0 g, 48 mmol, 83% yield); mp 96.0-97.5 °C (lit.12 94-96 °C).
10c. 1H NMR (CDCl3): δ 6.14 (second order “inverted
triplet”, 2H, H4 and H5), 7 4.23 (m, 1H, H3), 3.335 (td, J ) 10
and 4 Hz, 1H, H1), 3.225 (dd, J ) 10 and 6 Hz, 1H, H2), 3.26
(s, 3H, MeO), 2.71 (dd, J ) 16 and 6 Hz, 1H, H6), 2.67-2.46
(m, 1H, H6′). 13C NMR (CDCl3): δ 174.14 (C7), 171.15 (C8),
131.86 (C4), 126.58 (C5), 70.12 (C3), 56.58 (OCH3), 45.77 (C1),
36.48 (C2), 21.54 (C6).
1
11a . H NMR (CDCl3): δ 6.96 (d, J ) 9 Hz, 2H, H11′), 6.62
(d, J ) 9 Hz, 2H, H10′), 3.88 (s, 2H, H13). HRMS (DCI, CH4):
3-Tr im eth ylsilyloxy-cis-1,2,3,6-tetr a h yd r op h th a lic An -
h yd r id e (10d ). The title compound was prepared by dissolving
equimolar amounts of 1-trimethylsilyloxy-1,3-butadiene (8d ,
mixture of cis and trans isomers, Aldrich, 8.54 g, 60 mmol)
and maleic anhydride (5.88 g, 60 mmol) in acetonitrile (100
mL). The reaction mixture was stirred at room temperature
for 24 h under nitrogen. The solvent was rotary evaporated to
give a light yellow liquid (14 g; 58 mmol; 97% yield). The
product was identified on the basis of its spectral data, which
compared favorably with those of the known 3-(tert-butyldi-
methylsilyloxy)-6-methyl analogue.15 Only one set of absorp-
calcd (C21H20N2O2, M+) 332.1525; obsd 332.1527.
12a . TMA: Tm ) 206.19 °C (confirmed by DSC). H NMR
1
(CDCl3): δ 7.255 and 7.155 (AA′BB′m, 8H, H10 and H11), 5.97
(m, 4H, H4 and H5), 4.01 (s, 2H, H13), 3.20 (m, 4H, H1 and H2),
2.70 (dm, 4H, H3 and H6); 2.30 (dm, 4H, H3′ and H6′). 13C NMR
(CDCl3): δ 179.33 (C7 and C8), 140.93 (C9), 130.28 (C12), 129.76
(C11), 127.91 (C4 and C5), 126.65 (C10), 41.20 (C13), 39.32 (C1
and C2), 23.89 (C3 and C6). IR (KBr): 1780 and 1709 (imide
CdO), 1625 (CdC), 1512 cm-1 (Ar). HRMS (DCI, CH4): calcd
(C29H26N2O4, M+) 466.1893; obsd 466.1901.
1
tions were observed in the H NMR, though two sets could be
11b. TMA Tm ) 90.79 °C (confirmed by DSC). 1H NMR
(CDCl3): δ 7.36-7.14 (m, 9H, aryl), 7.03 (d, J ) 8 Hz, 2H,
detected in the 13C NMR spectral data. On the basis of peak
heights, the two components, presumably 10d (cis) and 10d ′
(trans), are present in an approximate ratio of 95:5, respec-
tively.
H
11′), 6.46 (d, J ) 8 Hz, 2H, H10′), 6.14 (m, 2H, H4 and H5),
3.94 (m, 1H, H3), 3.85 (s, 2H, H13), 3.46 (dd, J ) 9 and 6 Hz,
1H, H2), 3.345 (td, J ) 9 and 2 Hz, 1H, H1), 2.99 (dm, J ) 14
Hz, 1H, H6), 2.565 (ddm, J ) 14 and 9 Hz, 1H, H6′). 13C NMR
(CDCl3): δ 178.83 (C7), 176.80 (C8), 141.76 (Cipso), 140.71 (C9′),
140.50 (C9), 138.31 ((C12′ and C12), 129.60 (C11′), 129.45 (C11),
128.98 (Co), 128.68 (Cm), 127.69 (C4), 127.63 (Cp), 127.04 (C5),
126.48 (C10), 126.42 (C10′), 44.72 (C2), 41.14 (C13), 40.92 (C3),
37.90 (C1) 21.69 (C6). IR (KBr): 3471 and 3358 (NH2), 1779
and 1711 (imide CdO), 1600 (CdC), 1513 cm-1 (Ar). HRMS
(DCI, CH4): calcd (C27H24N2O2, M+) 408.1838; obsd 408.1827.
1
10d . H NMR (CDCl3): δ 6.012 (ddd, J ) 10, 5, and 2 Hz,
1H, H4), 5.925 (ddd, J ) 10, 6, and 3 Hz, 1H, H5), 4.534 (dd, J
) 5 and 4 Hz, 1H, H3), 3.295 (td, J ) 10 and 5 Hz, 1H, H1),
3.065 (dd, J ) 10 and 4 Hz, 1H, H2), 2.638 (ddt, J ) 18, 5, and
2.5 Hz, 1H, H6), 2.442 (ddd, J ) 18, 10, and 5 Hz, 1H, H6′),
-0.019 (s, 9, Me3Si). 13C NMR (CDCl3): δ 174.45 (C7), 171.61
(C8), 130.19 (C4), 128.27 (C5), 62.59 (C3), 47.22 (C1), 35.98 (C2),
20.74 (C6), -0.22 (Me3Si). IR (KBr): 1867, 1779, and 1711
(anhydride), 1658 (CdC), 1252 (Si-C), 1056 and 846 cm-1 (Si-
O-C). HRMS (EI): calcd (C10H13O4Si, M+-CH3) 225.0583;
obsd 225.0563.
12b. DSC: Tm ) 87.1 °C. 1H NMR (CDCl3): δ 7.39-7.17
(m, 18H, aryl), 6.125 (m, 4H, H4 and H5), 4.13 (bs, 2H, H13),
4.03 (s, 2H, H3), 3.42 (dd, J ) 8 and 3 Hz, 2H, H2), 3.245 (ddd,
J ) 10, 8, and 2 Hz, 2H, H1), 2.55 (m, 4H, H6 and H6′). 13C
NMR (CDCl3): δ 179.17 (C7), 178.31 (C8), 141.76 (Cipso), 140.09
(C9), 130.54 (C4), 130.25 (C12), 129.82 (C0), 128.95 (C11), 127.63
(C5), 127.53 (Cm), 126.95 (Cp), 126.52 (C10), 46.79 (C2), 41.23
(C13), 40.06 (C3), 38.67 (C1) 23.79 (C6). IR (KBr): 1779 and 1711
(imide CdO), 1600, 1513 cm-1 (Ar). HRMS (DCI, CH4): calcd
(C41H34N2O4, M+) 618.2519; obsd 618.2538.
10d ′. 13C NMR (CDCl3): δ 174.14 (C7), 172.10 (C8), 130.96
(C4), 129.08 (C5), 62.81 (C3), 45.95 (C1), 36.88 (C2), 21.69 (C6),
1.91 (Me3Si).
3,6-Dip h en yl-cis-1,2,3,6-tetr a h yd r op h th a lic An h yd r id e
(10f). Although the literature13 suggests preparing 10f in a
Parr bomb, we found that this is unnecessary. The title
compound was prepared instead by refluxing equimolar
amounts of trans,trans-1,4-diphenyl-1,3-butadiene (8f, Aldrich,
8.25 g, 40 mmol) and maleic anhydride (4 g, 40 mmol) in
p-xylene (100 mL) for 48 h under nitrogen. On cooling a white
solid formed, which was filtered, washed with ether, and air-
dried to give 10f (10 g, 33 mmol, 82% yield); mp 205-209 °C.
Further recrystallization from chloroform yielded purified
product (8.2 g, 27 mmol, 67% overall yield); mp 215-217 °C
(lit.13 208-209 °C).
1
11c. H NMR (CDCl3): 6.98 (d, J ) 9 Hz, 2H, H11′), 6.62 (d,
J ) 9 Hz, 2H, H10′). HRMS (DCI, CH4): calcd (C22H22N2O3,
M+) 362.1630; obsd 362.1611.
12c. TMA: Tm ) 75.1 °C. 1H NMR (CDCl3): δ 7.27
(AA′BB′m, 4H, H11), 7.19 (AA′BB′m, 4H, H10), 6.17 (m, 4H, H4
and H5), 4.31 (m, 2H, H3), 4.03 (s, 2H, H13), 3.28 (s, 6H, OMe),
3.165 (dd, J ) 14 and 8 Hz, 2H, H2), 3.15 (m, 2H, H1), 2.665
(m, 4H, H6 and H6′). 13C NMR (CDCl3): δ 179.33 (C7), 176.40
(C8), 140.93 (C9), 132.23 (C4), 130.47 (C12), 129.82 (C11), 127.84
(C5), 126.76 (C10), 71.73 (C3), 56.70 (OMe), 45.31 (C2), 41.29
(C13), 36.76 (C1) 22.28 (C6). IR (KBr): 1779 and 1711 (imide
CdO), 1611 (CdC), 1513 cm-1 (Ar). HRMS (DCI, CH4): calcd
(C29H18N2O4, M+-2CH3OH-2H2) 458.1267; obsd 458.1260.
10f.13 1H NMR (CDCl3): δ 7.47-7.33 (m, 10H, aryl), 6.558
(s, 2H, H4 and H5), 3.846 (bd, J ) 4.8 Hz, 2H, H3, and H6s
double resonance studies reveal that the broadening stems
from small couplings with the vinyl and aromatic hydrogens),
3.740 (dd, J ) 4.8 and 2.4 Hz, 2H, H1, and H2). 13C NMR
(CDCl3): δ 169.69 (C7 and C8), 137.85 (Cipso), 132.05 (C4 and
C5), 128.72 (Cm), 128.63 (Co), 127.67 (Cp), 47.46 (C3 and C6),
41.18 (C1 and C2). IR (KBr): 1849, 1812, and 1773 (anhydride),
1656 cm-1 (CdC). HRMS (DCI, CH4): calcd (C20H16O3, M+)
304.1099; obsd 304.1070.
12d . 1H NMR (CDCl3): δ 7.19 (m, 4H, H11), 6.90 (m, 4H,
H
H
10), 5.90 (m, 4H, H4 and H5), 4.56 (m, 2H, H3), 3.71 (s, 2H,
13), 3.185 (td, J ) 10 and 5 Hz, 1H, H1), 2.955 (dd, J ) 10
and 4 Hz, 1H, H2), 2.59 (m, 1H, H6), 2.44 (m, 1H, H6′), -0.09
(s, 9, Me3Si).

1292 Meador et al.
Macromolecules, Vol. 37, No. 4, 2004
Ta ble 1. P r od u ct Distr ibu tion in th e Air Th er m olysis of a 2:1 Mixtu r e of An h yd r id es 10a -e a n d Meth ylen ed ia n ilin e (11:
Mon oim id e; 12: Bisim id e; 13: Cr oss-Lin k ed Str u ctu r e; 14: Ha lf-Ar om a tized P r od u ct; 15: F u lly Ar om a tized P r od u ct)
anhydride 10
a (3-H)
204 °C (1 h)
315 °C (30 min)
343 °C (50 min)
371 °C (30 min)
11a (2%), 12a (98%)
11a (0.5%), 12a (98%),
13a (1%), 14a (-), 15a (0.5%) (35%), 14a (29%), 15a (1%)
11b (16%), 12b (38%),
13b (-), 14b (14%),
15b (32%)
11a (-), 12a (35%), 13a
11a (-), 12a (11%), 13a
(40%), 14a (37%), 15a (12%)
11b (-), 12b (-), 13b (11%),
b (3-Ph)
11b (67%), 12b (33%)
11b (7%), 12b (12%), 13b
(6%), 14b (35%), 15b (32%), 14b (21%), 15b (56%),
21 (8%)
21 (12%)
c (3-OMe)
11c (<1%), 12c (>95%), 12a (12%), 13a (17%),
12a (-), 13a (19%),
12a (-), 13a (19%),
15a (5%)
14a (47%), 15a (24%)
12g (-), 13a (35%),
14a (13%), 15a (52%)
11e (6%), 12e (85%),
13e (-), 14e (-), 15e (9%)
11f (2%), 12f′ (88%),
13f (-),15f (10%)
14a (47%), 15a (34%)
14a (33%), 15a (49%)
a
a
d (3-OSiMe3) 12g (61%), 13a (4%),
14a (35%)
e (4-Me)
-
-
11e (7%), 12e (93%)
11e (-), 12e (34%), 13e
(28%), 14e (15%), 15e (23%) (31%), 14e (14%), 15e (51%)
11f (-), 12f′ (2%), 13f (-),
15f (86%), 21 (12%)
11e (-), 12e (4%), 13e
f (3,6-diPh)
11f (75%), 12f′ (25%)
11f (-), 12f′ (-), 13f (-),
15f (85%), 21 (15%)
a
This stage was skipped because all bisimide 12 had already reacted.
Ta ble 2. P r od u ct Distr ibu tion (%) in th e Th er m olysis of 12a in Air a n d In er t Atm osp h er e a t 343 °C for 3 h
12a
13a
14a
15a
conditions
under air
cyclohexene bisimide
cross-linked
half- aromatized
fully aromatized
0
7
42
77
40
16
18
in inert atmosphere
1
11e. H NMR (CDCl3): δ 6.96 (d, J ) 8 Hz, 2H, H11′), 6.62
(d, J ) 8 Hz, 2H, H10′), 3.88 (s, 2H, H13). HRMS (DCI, CH4):
calcd (C22H22N2O2, M+) 346.1681; obsd 346.1679.
IR (KBr): 1777 and 1718 (imide CdO), 1511 cm-1 (Ar). HRMS
(DCI, CH4): calcd (C29H18N2O4, M+) 458.1267; obsd 458.1264.
1
17a . H NMR (CDCl3): δ 8.005 (AA′BB′m, 4H, H3 and H6),
12e. 1H NMR (CDCl3): δ 7.25 (AA′BB′m, 4H, H11), and 7.14
(AA′BB′m, 4H, H10), 5.61 (m, 2H, H5), 4.01 (s, 2H, H13), 3.22
(m, 4H, H2 and H1), 2.62 (m, 2H, H6 or H6′), 2.28 (m, 4H, H3
and H6′ or H6), 1.76 (s, 6H, CH3). 13C NMR (CDCl3): δ 179.45
(C7), 179.20 (C8), 140.83 (C9), 136.57 (C4), 130.21 (C11), 129.66
(C10), 126.42 (C12), 120.15 (C5), 41.13 (C13), 39.68 (C2), 39.25
(C1), 28.91 (C3), 24.53 (C6), 23.51 (CH3). IR (KBr): 1779 and
1708 (imide CdO), 1611 (CdC), 1512 (Ar) cm-1. HRMS (DCI,
CH4): calcd (C31H30N2O4, M+) 494.2206; obsd 494.2182.
8.003 (d with hyperfine AA′XX′ splitting, J ) 8 Hz, 4H, H11),
7.837 (AA′BB′m, 4H, H4 and H5), 7.673 (d with hyperfine
AA′XX′ splitting, J ) 8 Hz, 4H, H10). 13C NMR (CDCl3): δ
194.70 (C13), 166.95 (C7 and C8), 136.62 (C9), 135.71 (C12),
134.85 (C4 and C5), 131.73 (C1 and C2), 130.99 (C11), 126.11
(C10), 124.11 (C3 and C6). IR (KBr): 1783 and 1712 (imide Cd
O), 1733 (CdO), 1513 cm-1 (Ar). HRMS (DCI, CH4): calcd
(C29H16N2O5, M+) 472.1059; obsd 472.1074.
21. TMA: Tm ) 90.19 °C (confirmed by DSC). 1H NMR
(CDCl3): δ 7.285 and 7.199 (ABq with additional hyperfine
AA′BB′ splitting, J ) 8.4 Hz, 4H each, H6 and H5 respectively),
4.024 (s, 2H, H8), 2.859 (s, 8H, H3). 13C NMR (CDCl3): δ 176.24
(C2), 140.90 (C7), 130.02 (C4), 129.76 (C6), 126.48 (C5), 40.07
(C8), 28.34 (C3). IR (KBr): 1776 and 1706 (imide CdO), 1512
cm-1 (Ar). HRMS (DCI, CH4): calcd (C21H18N2O4, M+) 362.1267;
obsd 362.1300.
11f.16 TMA Tm ) 131.09 °C (confirmed by DSC). 1H NMR
(CDCl3): δ 7.48-7.20 (m, 14H, aryl), 7.044 (d, J ) 8.4 Hz, 2H,
H
11′), 6.860 (d, J ) 8.4 Hz, 2H, H10′), 6.496 (s, 2H, H4 and H5),
3.882 (bd, J ) 4.9 Hz, 2H, H3 and H6), 3.849 (s, 2H, H13), 3.612
(dd, J ) 4.9 and 2.2 Hz, 2H, H1 and H2). 13C NMR (CDCl3): δ
174.89 (C7 and C8), 143.49 (Cipso), 140.50 (C9′), 139.10 (C12′ and
C
12), 137.42 (C9), 131.27 (C4 and C5), 129.38 (C11′), 128. 89 (Co),
Th er m olysis of Bisim id es 12. A 2:1 mixture of anhydride
10 and MDA was heated for 1 h at 204 °C (400 °F; generally
yielding cyclohexene bisimides 12, as described above), then
at 315 °C (600 °F) for 30 min, at 340 °C (650 °F) for 50 min,
and finally at 371 °C (700 °F) for 30 min. After each step, the
reaction product was cooled, weighed, and ground to a powder,
and a sample was removed for spectroscopic analysis. Product
yields for thermolyses carried out in air and in inert atmo-
sphere (after deoxygenation via five vacuum/argon cycles) are
summarized in Tables 1 and 2, respectively. In air, both the
cross-linking of 12 to 13, and its aromatization to 15, were
observed. The cross-linking can be readily detected by the
appearance of aliphatic multiplets in region of 1.5-2.5 ppm.
On the other hand, the phthalimides formed on aromatization
have a pair of distinctive AA′BB′ multiplets located downfield
in the region 7.6-8.0 ppm and separated by ca. 0.2 ppm. The
HRMS data confirm the formation of half (14) and fully (15)
aromatized imides. In the case of the thermolysis of 12c, only
13a , 14a , and 15a were observed. For 12d , 12g was also
detected, in addition to 13a , 14a , and 15a (see Table 1). In
the case of the analogues 12e, 13e, and 15e, the respective
vinyl, alkyl, and aromatic methyls are easily discernible at
1.76, ca. 0.95 (vide infra), and 2.54 ppm. What is more, as
expected, there are three methyl groups observed for the cross-
linked 13e, each corresponding to one of the following iso-
mers: head-to-head, head-to-tail, and tail-to-tail.
128.81 (C11), 128.32 (Cm), 127.99 (Cp), 127.13 (C10), 126.20 (C10′),
46.41 (C1 and C2), 41.60 (C3 and C6), 40.87 (C13). IR (KBr): 3414
and 3379 (NH2), 1782 and 1714 (imide CdO), 1682 (CdC),
1597, 1513 cm-1 (Ar). HRMS (DCI, CH4): calcd (C33H28N2O2,
M+) 484.2151; obsd 484.2160.
12f′. TMA Tm ) 117.87 °C (confirmed by DSC). 1H NMR
(CDCl3): δ 7.48-7.10 (m, 28H, aryl), 6.029 (s, 4H, H4 and H5),
4.012 (s, 2H, H13), 3.882 (bd, J ) 4.0 Hz, 4H, H3 and H6), 3.202
(dd, J ) 4.0 and 1.6 Hz, 4H, H1 and H2). 13C NMR (CDCl3): δ
176.77 (C7 and C8), 143.42 (Cipso), 140.68 (C9), 129.98 (C12),
129.85 (C4 and C5), 129.62 (C11), 128. 80 (Co), 127.98 (Cm),
127.09 (Cp), 126.16 (C10), 45.98 (C1 and C2), 41.02 (C13), 39.75
(C3 and C6). IR (KBr): 3414 and 3379 (NH2), 1782 and 1714
(imide CdO), 1682 (CdC), 1597, 1513 cm-1 (Ar). HRMS (DCI,
CH4): calcd (C53H42N2O4, M+) 770.3145; obsd 770.3135.
12g. 1H NMR (CDCl3): δ 7.3-7.1 (m, 8H, H10 and H11), 6.12
(m, 2H, H4), 5.94 (m, 2H, H5), 4.53 (m, 2H, H3), 4.01 (s, 2H,
H
13), 3.92 (bs, 1H, OH), 3.45 (dd, J ) 10 and 6 Hz, 2H, H2),
3.30 (m, 2H, H1), 2.60 (m, 4H, H6 and H6′). 13C NMR (CDCl3):
δ 178.90 (C7 and C8), 141.30 (C9), 134.60 (C4), 131.11 (C12),
129.91 (C11), 126.89 (C5), 126.61 (C10), 66.36 (C3), 44.53 (C2),
41.29 (C13), 38.24 (C1) 23.88 (C6). IR (KBr): 3467 (OH), 1774
and 1708 (imide CdO), 1612 (CdC), 1512 cm-1 (Ar). HRMS
(DCI, CH4): calcd (C29H26N2O6, M+) 498.1791; obsd 498.1758.
15a . 1H NMR (CDCl3): δ 7.96 (AA′BB′m, 4H, H3 and H6)
and 7.79 (AA′BB′m 4H, H4 and H5), 7.37 (s, 8H, H10 and H11),
4.10 (s, 2H, H13). 13C NMR (CDCl3): δ 167.28 (C7 and C8),
140.46 (C9), 134.32 (C4 and C5), 131.79 (C12), 129.87 (C1 and
C2), 129.72 (C11), 126.60 (C10), 123.70 (C3 and C6), 41.20 (C13).
1
13a . H NMR (CDCl3): δ 4.02 (s, 2H, H13).
14a . 1H NMR (CDCl3): δ 4.05 (s, 2H, H13). HRMS (DCI,
CH4): calcd (C29H22N2O4, M+) 462.1580; obsd 462.1557.
15a . As above.

Macromolecules, Vol. 37, No. 4, 2004
Tetrahydrophthalic Anhydrides as End Caps 1293
1
13b. H NMR (CDCl3): δ 3.89 (s, 2H, H13).
1
14b. H NMR (CDCl3): δ 4.00 (s, 2H, H13); parent peak not
observed in HRMS.
15b.17 1H NMR (CDCl3): δ 7.95 (dd, J ) 4 and 1 Hz, 2H,
H6), 7.80 (t, J ) 4 Hz, 2H, H5), 7.70 (dd, J ) 4 and 1 Hz, 2H,
H4), 7.62-7.52 (m, 4H, pendant Ar), 7.52-7.38 (m, 6H,
pendant Ar), 7.38-7.10 (m, 8H, Ar), 4.03 (s, 2H, H13). 13C NMR
(CDCl3): δ 166.91 (C7), 166.81 (C8), 141.42 (C3), 140.37 (ipso),
140.14 (C2), 136.35 (C4), 135.97 (C12), 134.08 (C5), 132.84 (C9),
129.53 (C11), 129.39 (Co), 128.76 (Cp), 128.02 (Cm), 126.98 (C1),
126.62 (C10), 122.79 (C6), 41.03 (C13). HRMS (DCI, CH4): calcd
(C41H26N2O4, M+) 610.1893; obsd 610.1840.
13e. 1H NMR (CDCl3): δ 0.99, 0.96, and 0.93 (each s, CH3),
4.02 (s, 2H, H13).
14e. 1H NMR (CDCl3): δ 4.08 (s, 2H, H13). HRMS (DCI,
CH4): calcd (C31H26N2O4, M+) 490.1893; obsd 490.1888.
15e. 1H NMR (CDCl3): δ 7.82 (d, J ) 4 Hz, 2H, H6), 7.74 (s,
2H, H3), 7.70 (d, J ) 4 Hz, 2H, H5), 4.05 (s, 2H, H13), 2.54 (s,
6H, CH3). HRMS (DCI, CH4): calcd (C31H22N2O4, M+) 486.1580;
obsd 486.1521.
15f. 1H NMR (CDCl3): δ 7.712 (s, 4H, H6), 7.63-7.55 (m,
8H, pendant Ar), 7.50-7.40 (m, 12H, pendant Ar), 7.40-7.00
(m, 8H, Ar), 4.02 (s, 2H, H13). 13C NMR (CDCl3): δ 166.91 (C7
and C8), 141.42 (C3 and C6), 140.34 (ipso), 140.31 (C1 and C2),
140.01 (C9), 136.33 (C4 and C5), 136.18 (C12), 129.46 (Co), 128.71
(Cp), 128.71 (C11), 128.02 (Cm), 126.72 (C10), 41.03 (C13). HRMS
(DCI, CH4): calcd (C53H34N2O4, M+) 762.2519; obsd 762.2534.
F igu r e 3. Weight loss on cure and thermolysis of bisimides
12 formed from 1:2 mixtures of MDA and anhydrides 10.
Imidization only accounts for between 4 and 7% of the weight
loss observed.
above rule were the 3-phenyl (10b), 3,6-diphenyl (10f),
and 3-silyloxy (10d ) derivatives. In the case of 10b and
10f, the ratio of monoimide:bisimide (11:12) was 2:1 and
3:1, respectively, with bisimidization presumably slowed
by steric hindrance. The bisimides 12b and 12f were
the primary product when the reaction mixture was
heated further to 315 °C (see Table 1). Noteworthy is
the fact that while the vinyl and bridgehead hydrogens
in 3,3-diphenyl anhydride (10f) and monoimide (11f)
resonate at ca. 6.5 and 3.7 ppm, respectively, they are
observed ca. 0.5 ppm upfield at 6.05 and 3.20 ppm in
the corresponding bisimide 12f. We believe that this
corresponds to a conversion of the highly congested cis,
cis, cis 12f (with the phenyls and anhydride rings on
the same face) to the less hindered trans, cis, trans
isomer 12f ′ (eq 2); this isomerization will, in turn, affect
the diamagnetic anisotropy felt by the aforementioned
hydrogens.20 Facile high-temperature (315 °C) enoliza-
tion presumably mediates this isomerization.
Resu lts a n d Discu ssion
We have synthesized several substituted tetrahydro-
phthalic anhydrides for possible use as end caps18
s
including the 3-phenyl (10b), 3-methoxy (10c), 3-tri-
methylsilyloxy (10d ), and 3,6-diphenyl (10f) analoguess
via the Diels-Alder addition of the corresponding
butadienes 8 and maleic anhydride 9 (eq 1). The
butadienes are commercially available, with the excep-
tion of trans-1-phenylbutadiene, which was prepared
according to the procedure of Grummit and Becker.14
As shown in eq 1, when substituents R and R′′ lie trans
to the diene moiety in 8, they end up in 10 (and the
corresponding imides 11 and 12) aligned predominantly,
if not exclusively, cis to the anhydride (or imide) moiety.
The hydrogens at C3, C2, C1, and C6 are aligned cis, cis,
cis to each other. This is a result of the preferred endo
addition in Diels-Alder reactions.13,19 Two additional
tetrahydrophthalic anhydrides, the 3-hydro (10a , Ald-
rich) and 4-methyl (10e, TCP-EP) analogues, are com-
mercially available.
Anhydrides 10 and methylenedianiline in a 2:1 ratio
were then ground up together and heated at 204 °C (400
°F) for 1 h. As a rule and as determined by NMR, the
primary product at this stage was bisimide 12, which
was accompanied by a small amount (<2%) of the
intermediate monoimide 11 (eq 1). The spectrum of the
latter is quite distinctive because of the AA′XX′ pattern
of the H-10′ and H-11′ aromatic protons at ca. 6.6 and
7.0 ppm, respectively. Three notable exceptions to the
In the case of the 3-silyloxy derivative 10d , heating
at 204 °C resulted in a 34% weight loss (see Figure 3),
accompanied by the disappearance of the distinctive
trimethylsilyloxy peak in the NMR of the residual solid.
The NMR and HRMS data strongly indicate that the

1294 Meador et al.
Macromolecules, Vol. 37, No. 4, 2004
product obtained is the 3-hydroxybisimide 12g. DSC and
TGA/MS/IR studies suggest that imidization and hy-
drolysis occur in tandem at ca. 155 °C, accompanied by
the formation and loss of hexamethyldisiloxane (com-
mercially available from Aldrich). Theoretical weight
loss for both water due to imidization and hexamethyl-
disiloxane is 28%. Additional weight loss is due to
aromatization of some 35% of the 3-hydroxybisimide 12g
to half-aromatized product 14.
The polyimide model compounds (n ) 0) were heated
gradually to 370 °C and the thermal-oxidative trans-
formations followed by NMR. Product yields are sum-
marized in Table 1. Figure 3 presents the weight loss
data. Particularly noteworthy is the observation that
the tetrahydrophthalic end caps not only experience the
expected cross-linking yielding 13, but aromatize as well
to the phthalic analogue 15 (eq 3).
tions at ca. 4 ppm but three! Since this third peak
increases initially and then decreases, we believe it
corresponds to the initially formed nonsymmetrical half-
aromatized bisimide 14 (appearing at 4.05 ppm for 14a ;
eq 2). Indeed, HRMS data confirm the presence of 14
in the thermolysis product mixtures (see Experimental
Section). We additionally believe that the methylene for
the cross-linked product 13 appears very close (ca. 0.01
ppm downfield at 4.02 ppm for 13a ) to the chemical shift
of 12. Indeed, during the thermal treatment, the me-
thylene peak of 12 broadens and shifts ca. 0.01 ppm
downfield with the concomitant decrease in the olefinic
hydrogens at ca. 6 ppm.
Integration of the various peaks at ca. 4 ppm was used
to calculate the product distribution appearing in Table
1. These calculations were confirmed by use of the other
distinctive resonances (olefinic, phthalic, etc.) mentioned
above. It must be noted, however, that there may well
be some small amount of aromatic product formed at
elevated temperatures arising from already cross-linked
structures which are expected to appear under the broad
aromatic multiplet appearing between 7.4 and 7.2 ppm.
As noted in the Introduction, this is actually the desired
scenario, where cross-linking occurs through the double
bond of the end cap, followed by preferentially path B
degradation on further aging. We simply cannot distin-
guish in the peaks at ca. 4 ppm between cross-linked
aromatic structures and 15.
The cross-linking can be readily detected by the
appearance of multiplets in the region of 1.5-2.2 ppm.
(Interestingly, Table 1 reveals that cross-linking begins
at temperatures substantially lower than the 415 °C
indicated by St. Clair and St. Clair.10) On the other
hand, the phthalimides formed on aromatization have
a pair of distinctive AA′BB′ multiplets located downfield
in the region 7.6-8.0 ppm and separated by ca. 0.2 ppm.
An authentic sample of 15a was prepared, as above, by
grinding together a 2:1 ratio of phthalic anhydride and
methylenedianiline (16a ) and then heating the mixture
at 204 °C for 1 h (eq 4). (For the purpose of comparison,
the phthalic bisimide of diaminobenzophenone (17) was
similarly prepared and found to be absent from the
product mixture.)
Such oxidative aromatizations are precedented.23 In
the case of 12a , 12b, and 12e, this presumably involves
the free radical hydroperoxidation of the cylohexene ring
at the allylic positions followed by elimination of H2O2
(eq 6).
In the case of the 4-methyl analogues 12e, 13e, and
15e, the respective vinyl, alkyl, and aromatic methyls
are easily discernible at 1.76, ca. 0.96 (vide infra), and
2.54 ppm. What is more, for the cross-linked 13e, there
are three methyl groups observed (at 0.99, 0.96, and 0.93
ppm in a ratio of 3:6:2), presumably21 corresponding to
head-to-head, head-to-tail, and tail-to-tail isomers (eq
5).
It is known that the central C-13 methylene group of
MDA is amazingly sensitive to the nature of the
substituents attached to the amines on either side.22
Thus, the chemical shift of the methylene in tetrahy-
drophthalimide 12 (4.01 ppm for 12a ) is different from
bisphthalimide 15 (4.10 ppm for 15a ). However, we were
surprised to see on heating not two methylene absorp-
When R is an oxy linkage, as in the 3-methoxy
analogue 12c and the 3-hydroxy analogue 12g (obtained
from trimethylsiloxy 10d ), aromatization is accompa-
nied and presumably initiated by elimination of the
ether substituent, ultimately yielding 15a (eq 7). Theo-
retical weight loss for loss of the methoxy groups is
approximately 12%. [As noted previously, the silyl ether
12d is first hydrolyzed at ca. 155 °C to the hydroxy
group (i.e., to 12g), which upon further heating is
converted to 13a and 15a .]
Consistent with the difference in the mechanism is
the observation that aromatization is not observed until
315 °C in the case of 12a , 12b, 12e, and 12f, but

Macromolecules, Vol. 37, No. 4, 2004
Tetrahydrophthalic Anhydrides as End Caps 1295
aromatic products are observed even at 204 °C for 12c
and 12g. The elimination of the methoxy group in
heating the 12c to 315 °C is reflected in the relatively
large weight loss (12%; see Figure 3) observed at this
stage.
In the thermolysis of the phenylated bisimides 12b
and 12f, aromatization at 343 °C is accompanied by
substantial weight loss and the formation of bissuccin-
imide 21 (Table 1 and Figure 3 and eq 8). An authentic
sample of 21 was prepared, as above, by grinding
together a 2:1 ratio of succinic anhydride and methyl-
enedianiline (16a ) and then heating the mixture at 204
°C for 1 h. The bissuccinimide 21 is readily identified
in the reaction mixture by the succinic ring C3 meth-
ylene which appears as a singlet at 2.86 ppm (1H NMR)
and at 28.34 ppm (13C NMR). In addition, the noncon-
jugated imide carbonyl at C2 appears at 176.24 ppm
(13C). We attribute the formation of 21 to a retro-Diels-
Alder reaction in which the resulting butadienes 8b or
8f are volatilized at the elevated temperature.24 The
concomitantly formed bismaleimide 20 is presumably
reduced by bisimide 12, which is undergoing aromati-
zation at the same time to 15.
F igu r e 4. Percent cross-linking (above) and degree of aro-
matization (below) in the air thermolysis of substituted bis-
imides 12 formed from 1:2 mixtures of MDA and 1,2,3,6-
tetrahydrophthalic anhydrides 10a -f.
that substitution on the double bond in the latter case
slows cross-linking slightly and somewhat favors aro-
matization, beginning even as low as 315 °C. We note
as well that higher temperatures favor aromatization.
The data in Table 1 reveal that in the case of 12a at
343 °C the ratio of cross-linking to degree of aromati-
zation [13a :(14/2 + 15)] is ca. 2:1; however, at 371 °C
this drops to 4:3. In the case of 12e, this ratio drops
from 1:1 to 1:2.
Turning now to the phenylated analogues, we observe
that cross-linking is partially inhibited in the case of
the 3-phenylated 12b and totally inhibited in the 3,6-
diphenyl analogue 12f. Presumably, the cause is steric
in nature. In the case of 12b, we again see a rise in the
rate of aromatization as we go from 343 to 371 °C. For
12f, with cross-linking inhibited, aromatization is es-
sentially complete by 343 °C.
We believe that it is this substantial percentage of
aromatic product, with the concomitant lowering of the
relative amount of cross-linking, which is responsible
for both the improved TOS of tetrahydrophthalic end-
capped polyimides and their substantial frangibility.9,10
If the mechanism for aromatization is an oxidative
one, then thermolysis of the tetrahydrophthalimides 12
under inert atmosphere should lower the amount of
aromatization dramatically. This is indeed the case, as
shown in Table 2. Two samples (0.5 g) of powdered 12a
were thermolyzed at 343 °C for 3 h: prior to thermoly-
sis, one of the samples was sealed under vacuum in a
sealed tube after deoxygenating the powder via five
vacuum/argon cycles; the second sample was left open
Figure 4 graphs the effect of temperature on the
amount of cross-linking and the degree of aromatization
(defined as the percent of fully aromatized product plus
half the percent of half-aromatized product) in the air
thermolysis of the various derivatives of 12. This figure
highlights several interesting observations.
Outstanding is the rapid rate of aromatization and
cross-linking observed for the trimethylsilyloxy case
10d ; indeed, by 315 °C, no unreacted bisimide 12
remains. As discussed above, by 200 °C in this system,
the bisimide 12d has been converted to the hydroxy
analogue 12g, and it is in fact the high reactivity of the
latter that is noteworthy.
Comparing the parent tetrahydrophthalic bisimide
12a with that of the 4-methyl analogue 12e, we find

1296 Meador et al.
Macromolecules, Vol. 37, No. 4, 2004
(11) Smissman, E. E.; Murray, R. J .; McChesney, J . D.; Houston,
L. L.; Pazdernick, T. L. J . Med. Chem. 1976, 19, 148-153.
(12) Mikhael, M. G.; Padias, A. B.; Hall, J r., H. K. Macromolecules
1993, 26, 4100-4104.
(13) Ballistreri, F. P.; Maccarone, E.; Perrini, G.; Tomaselli, G.
A.; Torre, M. J . Chem. Soc., Perkin Trans. 2 1982, 273-277
and references therein.
to air. In the case of the sample in inert atmosphere,
aromatization was substantially inhibited; thus, the
degree of aromatization dropped from 38% in air down
to 8%. At the same time, the ratio cross-linking to the
degree of aromatization rose from ca. 1:1 in air down to
almost 10:1. It should be noted that despite the exclu-
sion of oxygen, some residual aromatization is, never-
theless observed; it presumably results from radical-
induced disproportionation.
(14) Grummit, O.; Becker, E. I. Org. Synth. 1963, Collect. Vol.
IV, 771.
(15) Cameron, D. W.; Heisey, R. M. Aust. J . Chem. 1995, 48, 185-
198 (compound 7 therein).
We are currently investigating the thermal-oxidative
stability of either tetrahydrophthalic end-capped poly-
imides processed under an inert atmosphere or those
in which aromatization is inhibited by the substitution
pattern on the end cap.
(16) Similar NMR data have recently been reported for the
tridecafluorohexylphenyl analogue. See the Supporting In-
formation for compound 30 in: Werner, S.; Curran, D. P. Org.
Lett. 2003, 5, 3293-3296.
(17) The spectrum is very close to the reported 1H NMR data for
3-phenylphthalic anhydride (4-phenylisobenzofuran-1,3-di-
one). See: Brown, R. F. C.; Choi, N.; Eastwood, N. C. Aust.
J . Chem. 2000, 53, 161-166 (compound 13 therein).
(18) Meador, M. A. B. U.S. Patent Number 6,274,699, Aug 14,
2001.
Ack n ow led gm en t. We thank Daniel Scheiman of
QSS, Inc., for performing thermal analysis and FT-IR
and Linda Ingraham of the Ohio Aerospace Institute
for assistance with thermolysis runs. We also thank the
High Operating Temperature Propulsion Components
(HOT_PC) Program at NASA Glenn for their kind
support.
(19) March, J . Advanced Organic ChemistrysReactions, Mecha-
nisms and Structure, 3rd ed.; Wiley: New York, 1985; p 748.
(20) We thank Dr. Hugo E. Gottlieb (Magnetic Resonance Labora-
tory, Bar Ilan University) for this suggestion.
(21) Based on Bio-Rad ChemWindow 6.5 1H NMR calculations.
(22) Cf.: Milhourat-Hammadi, A.; Chayrigues, H.; Merienne, C.;
Gaudemer, A. J . Polym. Sci., Part A: Polym. Chem. 1994,
32, 203-217.
(23) (a) van Tamelen, E. E.; Hildahl, G. T. J . Am. Chem. Soc. 1956,
78, 4405-4412. (b) Howe, R.; McQuillan, F. J . J . Chem. Soc.
1958, 1513-1518. (c) Crowshaw, K.; Newstead, R. C.; Roger,
N. A. J . Tetrahedron Lett. 1964, 5, 2307-2311ssee Frimer,
A. A. In The Chemistry of Enones, Part 2; Patai, S., Rap-
poport, Z., Eds.; Wiley: Chichester, 1989; pp 781-921, 837-
838. (d) Howard, J . A.; Ingold, K. U. Can. J . Chem. 1967, 45,
785-792. (e) Buchi, G.; Pickenhagen, W.; Wuest, J . Org.
Chem. 1972, 37, 4192-4193. (f) Hendry, D. G.; Schuetzle, D.
J . Am. Chem. Soc. 1975, 97, 7123-7127.
(24) The boiling point of trans-1-phenyl-1,3-butadiene (8b) is
reported14 to be 80 °C at 8 mm. This corresponds (nomograph)
to a boiling point of ca. 220 °C at 760 mm. The boiling point
of commercial (Aldrich) trans,trans-1,4-phenyl-1,3-butadiene
(8f) is 350 °C. The TGA of the 2:1 mixture of 8f and MDA
heated to 204 °C shows a Td at 239 °C (corresponding to
imidization) and 352 °C (corresponding to retro-Diels-Alder),
with onset in both cases ca. 20 °C earlier. DSC confirms these
transitions.
Refer en ces a n d Notes
(1) Serafini, T. T.; Delvigs, P.; Lightsey, G. R. J . Appl. Polym.
Sci. 1972, 16, 905.
(2) Meador, M. A. B.; J ohnston, J . C.; Cavano, P. J . Macromol-
ecules 1997, 30, 515-19.
(3) See: Meador, M. A. Annu. Rev. Mater. Sci. 1998, 28, 599-
630.
(4) Meador, M. A. B.; Lowell, C. E.; Cavano, P. J .; Herrera-Fierro,
P. High Perform. Polym. 1996, 8, 363-379.
(5) Meador, M. A. B.; J ohnston, J . C.; Cavano, P. J .; Frimer, A.
A. Macromolecules 1997, 30, 3215-3223.
(6) Meador, M. A. B.; J ohnston, J . C.; Cavano, P. J .; Frimer, A.
A.; Gilinsky-Sharon, P. Macromolecules 1999, 32, 5532-5538.
(7) Meador, M. A. B.; J ohnston, J . C.; Hardy-Green, D.; Frimer,
A. A.; Gilinsky-Sharon, P.; Gottlieb, H. E. Chem. Mater. 2001,
13, 2649-2655.
(8) Meador, M. A. B.; Frimer, A. A. U.S. Patent Number 6,303,
744, Oct 16, 2001.
(9) Alston, W. B. NASA Glenn Research Center, personal com-
munication, Aug 16, 2002.
(10) St. Clair, A. K.; St. Clair, T. L. Polym. Eng. Sci. 1982 22,
9-14.
MA0353078