Importance of π-stacking in photoreactivity of aryl benzobisoxazole and aryl benzobisthiazole compounds

Article abstract of DOI:10.1021/ma021045i

Aryl benzobisoxazole and aryl benzobisthiazole compounds in the solid state and in solution have completely different photoreactivity. In the solid state, intermolecular π-stacking interactions between these molecules lead to excimer formation. The excime

Full text of DOI:10.1021/ma021045i

Macromolecules 2003, 36, 4699-4708
4699
Importance of π-Stacking in Photoreactivity of Aryl Benzobisoxazole and
Aryl Benzobisthiazole Compounds
Yin g-Hu n g So,*,† Steve J . Ma r tin ,† Br u ce Bell,‡ Cu r tis D. P feiffer ,‡
Rich a r d M. Va n Effen ,‡ Br itton L. Rom a in ,† a n d Steve M. Lefk ow itz‡
Advanced Electronic Materials Department, 1712 Building, and Analytical Sciences Laboratory,
1897 Building, The Dow Chemical Company, Midland, Michigan 48674
Received J uly 3, 2002; Revised Manuscript Received April 15, 2003
ABSTRACT: Aryl benzobisoxazole and aryl benzobisthiazole compounds in the solid state and in solution
have completely different photoreactivity. In the solid state, intermolecular π-stacking interactions between
these molecules lead to excimer formation. The excimer undergoes a photoinduced electron transfer to
generate an ion radical pair. In the presence of oxygen, the anion radical transfers an electron to molecular
oxygen to form superoxide. The cation radical undergoes bond cleavage followed by hydrogen abstraction
and other reactions to form benzobisoxazoles (or benzobisthiazoles), benzonitriles, and benzamides. In
solution, aryl benzobisoxazole and aryl benzobisthiazole molecules cannot readily π-stack, and therefore
photoinduced electron transfer and its subsequent reactions do not occur. The compounds in solution are
stable even after prolonged irradiation, suggesting that intermolecular π-stacking plays a very important
role in photoreactivity for these molecules. Reversible redox reagents, such as ferrocene compounds, are
found to retard the subsequent reactions from the photoinduced electron-transfer reaction and hence
improve the photostability of poly{(benzo[1,2-d:5,4-d′]bisoxazole-2,6-diyl)-1,4-phenylene}.
In tr od u ction
At high pressure, quadratic fluorescence quenching by
molecular oxygen was observed. The greater tendency
for 2 to π-stack at high pressure due to its planar
configuration contributed to the observed difference.
Solid-state morphology of molecular and polymeric
materials, in particular the formation of π-stacks, is a
topic of tremendous growing interest in biological,
organic, and polymer chemistry.1 Understanding the
effect of π-stacking on the photophysical and photo-
chemical properties of conjugated polymers and oligo-
mers is of theoretical as well as practical importance.
The planarity and rigidity of aryl benzobisoxazole and
aryl benzobisthiazole compounds and polymers enable
the molecules to form π-stacks readily. In a previous
publication, we reported dramatically, and sometimes
totally, different photophysical properties of these mol-
ecules in the solid state where they easily form π-stacks
and in solution where the molecules do not readily
π-stack.2 The findings, which were attributed to strong
π-stacking interaction in the solid state, are summarized
below.
(a) Large excitation and emission energy differences
were observed for planar p-phenylenebenzobisoxazole
and p-phenylenebenzobisthiazole model compounds and
polymers in solid state, but the same model compounds
in solution and those with side groups which hindered
the overall structural planarity showed significantly
smaller Stokes shifts.
(b) Only one emitting component was observed when
a benzobisoxazole model compound in solution was
excited. The same compound in the solid state showed
two emitting species.
(c) Fluorescence quenching by oxygen of 2,2′-(1,4-
phenylene)bis[6-(3,5-bis(1,1-dimethylethyl)phenyl]benzo-
[1,2-d;5,4-d′]bisoxazole (1), a compound with side groups
interrupting its planarity, followed the normal Stern-
Volmer quenching plot. The data for 2,6-bisphenylbenzo-
[1,2-d:5,4-d′]bisoxazole (2), which is planar,3 did not
follow a linear Stern-Volmer quenching relationship.
We report in this article that completely different
chemical reactions resulted when p-phenylenebenzo-
bisoxazole and p-phenylenebenzobisthiazole compounds
were photolyzed in the solid state, where the molecules
could stack, and in dilute solution, where π-stacking did
not happen. A photoinduced electron-transfer mecha-
nism is used to account for this dramatic difference.
Photodecomposition of the cation radical is retarded by
adding a reversible electron-transfer reagent to the
system.
Resu lts a n d Discu ssion
P h otolysis of Ar yl Ben zobisoxa zole a n d Ar yl
Ben zobisth ia zole Com p ou n d s in Solu tion . Aryl
benzobisoxazole and aryl benzobisthiazole compounds
have very low solubility in common organic solvents.
The very low solubility of these compounds imposes
† Advanced Electronic Materials.
‡ Analytical Sciences Laboratory.
10.1021/ma021045i CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/06/2003

4700 So et al.
Macromolecules, Vol. 36, No. 13, 2003
difficulties in following the disappearance of starting
materials and characterization and quantitation of
reaction products. Compound 1, 2,6-(4-tert-butyldiphen-
yl)benzo[1,2-d:5,4-d′]bisoxazole (3), and 2,6-(4-tert-bu-
tyldiphenyl)benzo[1,2-d:4,5-d′]bisthiazole (4), which have
tert-butyl groups to enhance their solubility, were
synthesized to allow the study of photolysis of aryl
benzobisoxazoles in solution.
MS. A large amount of CO2 was found in the serum vial
headspace of samples 1, 2, and 5 photolyzed in air.
GC-MS was also used to analyze photolyzed thin-film
samples of 2 and 5. GC-MS of photolyzed 2 showed
benzoic acid (0.1%), benzamide (1.0%), phthalimide
(0.1%), 9 (0.2%), and 2 (98.5%). Reconstructed total ion
chromatograms from the GC-MS analyses of an unex-
posed sample of 2 compared to a sample exposed in air
for 150 h are shown in Figure 1a,b. The mass spectra
of the products were compared with and were identical
to those of the control compounds. The photolyzed
sample of 5 contained 2-phenylbenzo[1,2-d:4,5-d′]bisthi-
azole (10) (0.2%) and 5 (99.8%). When exposure time
was increased from 150 to 430 h, the concentration of
10 and 5 was 0.5% and 99.5%, respectively. Formation
of 10 from photolysis of 5 was also confirmed with the
same procedure as for products from photolysis of 2. LC
analysis showed a lower concentration of remaining 2
and 5 than the concentration of 2 and 5 from the peak
areas of the total ion chromatograph. Results based on
reduction of LC peak area are the more accurate
reaction conversions. The large amount of CO2 found
in the serum vial headspace might be the reason for 2
and 5 not appearing in GC-MS analysis. Products from
photolysis of solid state 2 and 5 are shown in Scheme
1.
A 3.2 × 10-4 M solution of 3 or 4 in cyclohexane was
photolyzed under atmospheric conditions. After pho-
tolysis for 150 h, starting material was 100% recovered
as measured by weight. The solubility of 1 in cyclohex-
ane was so low that LC peak area integration and MS
were used to analyze the disappearance of starting
material and possible formation of photochemical prod-
ucts. After photolysis for 150 h, LC peak area integra-
tion showed solution concentration did not change, and
MS detected only 1 in the sample.
A methylene chloride solution (3.2 × 10-4 M) of 2 was
photolyzed for 100 h under atmospheric conditions. GC-
MS of the exposed solution showed 2 (96%) and chlori-
nated 2 (4%), which most likely was the product from
methylene chloride photodecomposition. Similar results
were observed with 2,6-bisphenylbenzo[1,2-d:4,5-d′]-
bisothiazole (5). When experiments were repeated with
glass flasks, chlorinated 2 and 5 were not detected. The
results suggested that aryl benzobisoxazole and aryl
benzobisthiazole compounds in solution are photolyti-
cally stable, which is consistent with the report by
Gu¨sten and co-workers that 2 and substituted 2 in
ethanol or dioxane were stable laser dyes.9
Analysis and confirmation of photolysis products of
6, 7, and 8 were performed with (a) controls, (b) chemical
ionization to confirm M+, (c) single ion profiles for
thermal separation of products, and (d) extraction and
analysis by GC-MS of lower molecular weight products.
Confirmation of photoreactivity of 6 is illustrated in
detail.
P h otolysis of Ar yl Ben zobisoxa zole a n d Ar yl
Ben zobisth ia zole Com p ou n d s in th e Solid Sta te.
For 1, 2, and 5, which have reasonable solubility in
chloroform, thin layers of about 0.3 µm thick were cast
from chloroform solutions on the inside wall of 100 mL
serum vials. The vials were either capped with air inside
or purged with nitrogen before capping and then irradi-
ated.2,6-Bis[4-(2-benzoxazoly)phenyl]benzo[1,2-d:5,4-d′]-
bisoxazole (6), 2,2′-(1,4-phenylene)bis(6-phenylbenzo-
[1,2-d:5,4-d′′]bisoxazole (7), and 2,6-bis[4-(2-benzothi-
azoyl)phenyl]benzo(1,2-d:4,5-d′′)bisthiazole (8) are to-
tally insoluble in nonacidic organic solvents. Thin films
were cast from TFA solutions in serum vials or depos-
ited on a microscope slide by sublimation.
Parts a and b of Figure 2 are the EI-MS of unexposed
6 and the EI-MS of 6 after it was exposed in air for 150
h. M+ and M2+ were confirmed by chemical ionization.
Figure 3 is the single ion profiles for thermal separation
of products off the direct exposure probe. Species with
specified m/z were detected as they came off of the probe.
Lower molecular weight products were extracted with
chloroform, and the chloroform solution was analyzed
by GC-MS. Figure 4 shows the total ion chromatogram
for such an extract. The mass spectra of the three eluted
Ta ble 1
exposure
sample
time (h)
peak area
%
Thin film samples of 1, 2, and 5 after photolysis in
the presence of oxygen were analyzed by LC. Table 1
shows the concentration of starting 1, 2, and 5 that
remained after photolysis in nitrogen or in air as
monitored by LC peak area. Reduction in starting
material peak area and the appearance of early-eluting,
lower molecular weight photoreaction products were
observed as compared to control samples. Gaseous
compounds in the vial headspace were analyzed by GC-
1 unexposed to light
1 exposed under nitrogen
1 exposed in air
2 unexposed to light
2 exposed under nitrogen
2 exposed in air
5 unexposed to light
5 exposed under nitrogen
5 exposed in air
0
150
150
0
150
150
0
430
150
430
546 000
547 000
497 000
435 000
434 000
396 000
591 000
592 000
579 000
549 000
100
100
91
100
100
91
100
100
98
5 exposed in air
93

Macromolecules, Vol. 36, No. 13, 2003
Aryl Benzobisoxazole and Aryl Benzobisthiazole 4701
F igu r e 1. (a) Total ion chromatograph of an unexposed sample of 2. (b) Total ion chromatograph of a sample of 2 photolyzed for
150 h in air.
compounds were identical to those of independently
synthesized 11, 12, and 13. Because of their lack of
solubility in chloroform and their low volatility, 4-(6-
phenylbenzo[1,2-d:5,4-d′]bisoxazol-2-yl)benzonitrile (14)
and 4-(6-phenylbenzo[1,2-d:5,4-d′]bisoxazol-2-yl)benzo-
amide (15) were detected by EI-MS as shown by single
ion profiles of thermally separated products, but they
did not show up in GC-MS analysis of the chloroform
extract.
The same methodology was used to analyze the
photolytic products from 7, 8, and 1. In the sample of
7, the compounds 9, 16, and 2-(4-(benzo[1,2-d:5,4-d′]-
bisoxazol-2-yl)phenyl)-6-phenylbenzo[1,2-d:5,4-d′]bisox-
azole (17) were found. For 8, the compounds 18, 2-(4-
(2-benzothiazolyl)phenyl)benzo[1,2-d:4,5-d′]bisthiazole
(19), 4-(6-(4-(2-benzothiazolyl)phenyl)benzo[1,2-d:4,5-d′]-
bisthiazol-2-yl)benzonitrile (20), and an unidentified
compound with m/z of 294 were detected after photoly-
sis. Sublimed 6, 7, and 8 produced the same products
as samples cast from TFA solution. 3,5-Bis(1,1-dimeth-
ylethyl)benzamide, 4-(6-(3,5-bis(1,1-dimethylethyl)phen-
yl)benzo[1,2-d:5,4-d′]bisoxazole (21), and 4-(6-(3,5-bis-
(1,1-dimethylethyl)phenyl)benzo[1,2-d:5,4-d′]bisoxazol-
2-yl)benzamide (22) were identified from photolyzed 1.
Figures 5 and 6 are the EI-MS of 7 and 8 photolyzed in
air. Products from photolysis of solid state 6, 7, 8, and
1 are shown in Scheme 2. The headspace of the serum
vials as analyzed by GC-MS contained a large amount
of CO2.

4702 So et al.
Macromolecules, Vol. 36, No. 13, 2003
Sch em e 1
When films of 1-8 were photolyzed in a nitrogen
atmosphere, only unreacted starting materials were
found by GC-MS or DEP-MS. None of the above-
mentioned products were detected. It has been reported
that physically damaged poly{(benzo[1,2-d:5;4-d′]bisox-
azole-2,6-diyl)-1,4-phenylene} (PBO) filaments exposed
to oxygen and light tend to lose a portion of their tensile
strength. Physical damage to the filament permits
oxygen to enter the otherwise oxygen impermeable
filament.8
P h otoin d u ced Electr on Tr a n sfer of Excim er s
Gen er a ted fr om π-Sta ck in g. We previously reported
that intermolecular π-stacking interactions occurred
between the planar aryl benzobisoxazole and aryl ben-
zobisthiazole compounds, leading to excimer formation
in solid state. In contrast, no excimer formation was
detected for these compounds in solution. Intermolecu-
lar π-stacking led to the disparity between photophysical
properties for aryl benzobisoxazole and aryl benzo-
bisthiazole compounds in solid state and in solution as
described in the Introduction section. These compounds
in the solid state underwent photoinduced electron
transfer, and in the presence of molecular oxygen, the
radical ion pair led to the generation of superoxide, as
shown in Scheme 3.13,14 Superoxide was trapped, and
the product was identified by ESR. The unstable cation
radical decomposes to generate a free radical and a
cation. Hydrogen abstraction by the free radical pro-
duces benzobisoxazoles and benzobisthiazoles. Oxygen,
superoxide, and moisture may react with the cation
radical to generate nitriles and amides.
The above results demonstrated the surprising but
completely different photoreactivity of aryl benzobisox-
azole and aryl benzobisthiazole compounds in solution
and in solid state. These compounds in solution are very
photolytically stable whereas in the solid state and in
the presence of oxygen, they are photolytically labile.
Although different photochemistry for organic com-
pounds in solid state and in solution has been a subject
of intense activity in the past 30 years,1 the totally
different photoreactivity of aryl benzobisoxazole and
aryl benzobisthiazole compounds in solid state and in
solution is one of the most dramatic.
Scheme 3 is also consistent with the fact that strong
ESR signals were detected for PBO and PBT fibers that
had been sealed in a vacuum for 40 days. Oxygen is not
required for the generation of the ion radicals. In the
absence of oxygen, the ion radical pair undergoes back
electron transfer. In solution, the planar molecules do
not π-stack, photoinduced electron transfer cannot oc-
cur, and the molecules are photolytically stable.
Cyclic Volta m m etr ic Stu d y of th e Red ox P r op -
er ties of P BO P olym er . The electrochemical reduction
of PBO and poly{(benzo[1,2-d:4,5-d′]bisthiazole-2,6-diyl)-
1,4-phenylene} (PBT) polymers is reversible whereas
oxidation is not reversible.10,11 Integration of the reduc-
tion peak current and calculation of the amount of
material on the electrode led to an n value of 0.9
electrons per repeating unit, indicating the formation
of a poly(anion radical). The formation of this reduced
material was accompanied by a color change from yellow
to black. The reduction peak and the color change could
be cycled repeatedly without apparent loss of film
mechanical integrity. The electrochemical oxidation of
PBO and PBT was totally irreversiblesone scan through
the oxidative process was sufficient to deactivate the
film electrochemically and destroy its mechanical in-
tegrity. Integration of the oxidation current yielded an
n value of 0.7 electrons per repeating unit. A poly(cation
radical) was formed which readily decomposed.
Effect of Rever sible Red ox Rea gen ts on P BO
P h otoch em istr y. According to Scheme 3, a reversible
redox reagent, which is able to transfer an electron to
the unstable cation radical to make a neutral molecule
and then pick up an electron from superoxide, should
be an efficient inhibitor of the photochemical reaction.
Chemical or anodic oxidation of ferrocene involves a one-
electron-transfer process to form ferricenium, which is
reversibly reduced back to ferrocene.15 Ferrocene com-
pounds such as 1,1′-ferrocene-dimethanol and ferrocen-
ecarboxylic acid were infiltrated into PBO fiber by
soaking the “wet” fiber in a 1 wt % aqueous solution of
the ferrocene compounds. Upon drying, the ferrocene
compounds were imbedded in the filaments. The amount
of iron in the fiber was typically 7000 ppm, which
indicated 3 wt % of ferrocene dimethanol. PBO fibers
containing ferrocene compounds showed much better
tensile strength retention upon photolysis in the pres-
ence of oxygen than PBO fiber without this redox
reagent.8 The results suggest that ferrocene inhibits the
photoreaction of PBO by back electron transfer, as
shown in Scheme 4.
ESR Sp ectr oscop y of P BO a n d P BT. Strong ESR
signals were observed for PBO, PBT, and their model
compounds.12 Oxygen was not required for the genera-
tion of the radicals.2 This observation was also reported
by DePra and co-workers.11 A PBO fiber sample, which
had been sealed in a vacuum for 40 days, showed strong
ESR signals. Formation of superoxide from the photoly-
sis of PBO, PBT, and their model compounds in the
presence of oxygen was detected by ESR spectroscopy
using 5,5-dimethyl-1-pyrroline N-oxide as a radical trap
as reported in ref 2.

Macromolecules, Vol. 36, No. 13, 2003
Aryl Benzobisoxazole and Aryl Benzobisthiazole 4703
F igu r e 2. (a) EI-MS of unexposed 6. (b) EI-MS of 6 photolyzed in air.
E ffect of Sin glet Oxygen on Ben zob isoxa zole
Com p ou n d s. Singlet oxygen was suggested to be a
reactive intermediate in the photoinduced degradation
of many polymers, including polyacetylene,16 poly(2,5-
bis(5,6-dihydrocholestanoxy)-1,4-phenylenevinylene) in
both solution and solid film,17 and poly(3-alkyl-
thiophenes).18 The effect of singlet oxygen on aryl
benzobisoxazole compounds was studied.

4704 So et al.
Macromolecules, Vol. 36, No. 13, 2003
F igu r e 3. Single ion profiles for thermal separation of products from photolyzed 6.
F igu r e 4. Total ion chromatograph of chloroform-extracted low molecular weight products from photolyzed 6.
Compound 2 in methylene chloride with methylene
blue as the sensitizer was irradiated for 100 h while
oxygen was bubbled through the solution. The starting
material was 100% recovered. 1,4-Bis(5-phenyloxazole-
ride.20 Aryl benzobisoxazole compounds appear to be
stable to singlet oxygen.
Exp er im en ta l Section
1
2-yl)benzene, which is known to react with O2, totally
Sta r tin g Ma ter ia ls a n d Con tr ol Sa m p les. The synthesis
and characterization of compounds 1-8 have been reported.2
2-Phenylbenzo[1,2-d:5,4-d′]bisoxazole (9) was made by stirring
2-phenyl(5-amino-6-benzoxazolol) in formic acid at reflux. The
reactant 2-phenyl(5-amino-6-benzoxazolol) was prepared by
decomposed under identical experimental conditions in
20 h.19 A thin film of 7, which was deposited on the
inside wall of a 100 mL flask, was inert to singlet oxygen
produced from hydrogen peroxide and sodium hypochlo-

Macromolecules, Vol. 36, No. 13, 2003
Aryl Benzobisoxazole and Aryl Benzobisthiazole 4705
F igu r e 5. EI-MS of 7 photolyzed in air.
F igu r e 6. EI-MS of 8 photolyzed in air.
reacting benzoic acid with excess 1,3-diamino-4,6-dihydroxy-
benzene dihydrogen chloride in poly(phosphoric acid). 4-(2-
Benzoxazolyl)benzonitrile (11) was made from o-aminophenol
and 4-cyanobenzoyl chloride.4 4-(2-Benzoxazolyl)benzamide

4706 So et al.
Macromolecules, Vol. 36, No. 13, 2003
Sch em e 2
(12) was made from 11 with 3% hydrogen peroxide and 25%
KOH.5 2-(4-(2-Benzoxazolyl)phenyl)benzo[1,2-d:5,4-d′]bisox-
azole (13) was made by refluxing 2-(4-(2-benzoxazolyl)phenyl)-
(5-amino-6-benzoxazolol) in formic acid.
sublimed film, a 10 mg sample was placed in a 100 mL, one-
necked flask containing a 1 cm × 5 cm microscope slide. The
flask was connected to a vacuum pump. The lower portion of
the flask was immersed in a sand bath at 350 °C. A thin film
of sublimed 6, 7, or 8 condensed on the upper portion of the
slide. Film thickness ranged from 0.2 to 0.3 µm as measured
by a profilometer, Alpha-Step 200 from Tencor Instruments.
P h otolysis P r oced u r es. Photolysis was performed with
either an apparatus manufactured by Oriel Corp., which was
equipped with
a 200 W medium-pressure mercury lamp
manufactured by Hanovia, or a Suntest CPS (Trademark of
Heraeus Co.). The Suntest used a filtered xenon lamp that
roughly mimics solar radiation in both intensity and wave-
length distribution. The output at the lamp source was 765
W/m2.
4-(6-Phenylbenzo(1,2-d:5,4-d′)bisoxazol-2-yl)benzonitrile (16)
was made from 2-phenyl(5-amino-6-benzoxazolol) and 4-cy-
anobenzoyl chloride. 4-(2-Benzothiazolyl)benzonitrile (18) was
synthesized from o-aminothiophenol and 4-cyanobenzoyl chlo-
ride.2
Sa m p le P r ep a r a tion . The compound of interest (2.00 mg)
was dissolved in chloroform or trifluoroacetic acid (TFA) in a
100 mL serum vial with inside surface area of approximately
100 cm2. A thin layer of the compound was formed on the
inside wall by rotating the vial to allow the solvent to
evaporate. Coating was visibly uniform. The thickness of the
layer was about 0.3 µm.6 Samples cast from TFA solutions
were heated in a vacuum oven to remove TFA. To prepare a
A 250 mL solution of the compound was placed in a quartz
or glass flask equipped with a stirring bar and a condenser.
The flask was 50 cm away from the Oriel apparatus, and a
water filter was placed between the flask and the light source.
Light intensity at the sample site was measured with a
Newport laser power meter model 820. The light intensity at
the site was about 200 W/m2. For photolysis of compounds in
solid state, the serum vial was placed horizontally in the
Suntest for radiation. The exposure time was 150 h unless
specified. For the sublimed sample, the microscope slide was
placed in the capped serum vial during photolysis. All solid-

Macromolecules, Vol. 36, No. 13, 2003
Aryl Benzobisoxazole and Aryl Benzobisthiazole 4707
Sch em e 3
Sch em e 4
serum vial at 80 °C was analyzed as the control. A computer
reconstructed total ion chromatograph showed the level of CO2
in the headspace over the photolyzed sample was about 30-
fold higher than that of air. Acetone and acetonitrile were
detected, but CO2 was the much bigger component in the
headspace. Sulfur dioxide was found in photolyzed aryl ben-
zobisthiazole compounds.
state photolysis experiments were performed with serum glass
vials.
Cyclic Volta m m etr ic Stu d y. Cyclic voltammetry was
performed with a Bioanalytical System (BAS, West Lafayette,
IN) model 110A electrochemical analyzer equipped with the
standard voltammetric cell stand. The working electrode was
a 3 mm diameter, glassy carbon disk, the reference electrode
was Ag/AgCl, and the counter electrode was a platinum wire.
A 0.14% solution of PBO polymer was prepared by dissolving
a known amount of fiber in methanesulfonic acid (MSA). A 1
µL aliquot of this solution containing 1.4 µg of PBO polymer
was applied to the electrode surface and spread around to
produce a uniform film of liquid. MSA was removed by rinsing
with a large volume of deionized water. Prior to voltammetric
An a lysis of P r od u cts fr om P h otoch em ica l Rea ction s.
(a ) Liqu id Ch r om a togr a p h y. Tetrahydrofuran (THF) (20
mL) was added to the vial, which was then placed on a bottle
roller until all visible residue was dissolved. A sample (10 µL,
0.1 mg/mL) was analyzed by a liquid chromatography (LC)
system consisting of a Perkin-Elmer model 410 gradient
pumping system, a Perkin-Elmer ISS-200 autosampler, and
a Perkin-Elmer 235 photodiode array UV detector. A Zorbax-
SB phenyl column (4.6 × 150 mm, 5 µm) was used for all
separations. The mobile phase was 62% THF and 38% water
with 0.02 M H3PO4 with a flow rate of 1.5 mL/min. The
detector wavelength was 270 or 330 nm. LC peak areas were
the average of three duplicated injections. The relative preci-
sion was 1% or better.
(b) Ga s Ch r om a togr a p h y-Ma ss Sp ectr oscop y. Chlo-
roform (10 mL) was added to the vial. A sample (1 µL) of the
filtered solution was injected onto a Finnigan SSQ700, SN
TS00417, quadrupole gas chromatography-mass spectroscopy
(GC-MS) system in the electron impact (EI) mode. The column
was a 15 m × 0.25 mm (0.25 µm film) DB-5 with a temperature
program of 80-320 °C at 16 °C/min. The MS was scanned from
35 to 650 atomic mass unit at 1 s/scan. Product identification
was performed by comparison of GC retention time and MS
fragmentation pattern of control compounds.
(c) Dir ect Exp osu r e P r obe (DEP ) Ma ss Sp ectr oscop y.
Control samples and photolyzed samples were dissolved in
TFA. A drop of the solution was placed on the DEP filament
probe. The current of the probe was ramped from 0 to 750 mA
at 500 mA/s. MS was run in EI and ammonia chemical
ionization modes.
(d ) An a lysis of Ser u m Via l Hea d sp a ce. A serum vial
containing a sample photolyzed for 150 h was warmed to 80
°C. A 5 mL aliquot of the headspace was injected onto the GC-
MS system in EI mode. The front 2 in. segment of the 25 m ×
0.25 mm (0.25 µm film) DB-FFAP (Durabond-Free Fatty-Acid
Phase) GC column from J &W Scientific was cooled with liquid
nitrogen to cryofocus the volatile organic compounds. The
liquid nitrogen cup was removed 2 min after sample injection.
The temperature program was 40 °C for 6 min and then 20
°C/min to 200 °C for 1 min. A room air blank from an empty
analysis, the electrode was placed in
a large volume of
acetonitrile to remove most of the water. The electrolyte was
0.1 M tetrabutylammonium perchlorate in acetonitrile, and
the scan rate was 10 mV/s. The number of electrons per repeat
unit was calculated from the Faraday equation since the
charge passed, the amount of material on the electrode, and
the molecular weight of the repeat unit were known.
Electr on Sp in Reson a n ce Sp ectr oscop y. Instrument
and measurement conditions for electron spin resonance (ESR)
spectroscopy were reported in ref 2.
Sta biliza tion of P BO F iber w ith F er r ocen e Com -
p ou n d s. PBO was synthesized from terephthalic acid and 1,3-
diamino-4,6-dihydroxybenzene in poly(phosphoric acid) (PPA).
Fiber was spun from the solution, and PPA was leached out
as the fiber was coagulated in water. The “wet” fiber was
soaked in a 1 wt % aqueous solution of 1,1′-ferrocene-
dimethanol (from Aldrich Chemical Co.) for 30 min. The fiber
was then air-dried and rinsed with water to remove the
ferrocene compound on the fiber surface. The processed fiber
contained 7000 ppm of iron as analyzed by neutron activation,
which indicated 3 wt % of ferrocene dimethanol.
The fiber was knitted into fabrics which were then deknitted
to introduce physical damage for oxygen penetration.7 The
fiber was exposed in the Suntest unit for 200 h. Tensile
strength of fiber samples was measured in accordance with
ASTM D-2101 on an Instron 4201 universal testing machine.
Samples embedded with ferrocene compounds showed sever-
alfold better tensile strength retention over control samples.8
Sin glet Oxygen on Ar yl Ben zoxa zole Com p ou n d s.
Compound 2 (0.312 g, 1.0 mmol) and methylene blue (0.94 g,

4708 So et al.
Macromolecules, Vol. 36, No. 13, 2003
A. Photochemical Solid to Solid Reactions. In Organic Mo-
lecular Photochemistry; Ramamurthy, V., Schanze, K. S.,
Eds.; Marcel Dekker: New Yrok, 1998; Vol. 2, pp 195-248.
(d) Photochemistry in Organized and Constrained Media;
Ramamurthy, V., Ed.; VCH: New York, 1991.
(2) So, Y. H.; Zaleski, J . M.; Murlick, C. L.; Ellaboudy, A.
Macromolecules 1996, 29, 2783-2795.
(3) Fratini and co-workers reported single-crystal studies of PBO
model compound (compound 2) and similar model compound
for PBT (compound 5). They reported 2 is planar (Fratini, A.
V.; Lenhert, P. G.; Resch, T. J .; Adams, W. W. In The
Materials Science and Engineering of Rigid-Rod Polymers;
Adams, W. W., Edy, R. K., McLemore, D. E., Eds.; Symposium
Proceedings, Vol. 134; Materials Research Society: Pitts-
burgh, 1989; pp 431-445), but 5 is not, with the torsion angle
of ∼23° between the phenyl and the benzobisthiazole moi-
eties: Wellman, M. W.; Adams, W. W.; Wolff, R. A.; Dudis,
D.; Wiff, D. R.; Fratini, A. V. Macromolecules 1981, 14, 935-
939.
0.25 mmol) were dissolved in 250 mL of methylene chloride
in a 500 mL, three-necked flask equipped with a condenser, a
glass tube extended to the bottom of the flask, and a stopper.
Oxygen was bubbled through the stirred solution. The experi-
ment was performed under room light. After 100 h, the
solution was analyzed by GC with 2-phenylbenzoxazole as the
internal standard. Analysis results showed 0.312 g of 2 (100%
recovery) in solution.
Compound 6 (3.0 mg) was dissolved in 10 mL of TFA and
deposited as a thin film on the inside wall of a 100 mL, round-
bottom flask. The thin film was treated with singlet oxygen
produced from H2O2 and NaClO as follows. Methanol (50 mL)
and 10 mL of H2O2 (30%) were added to the flask. The solution
was stirred. Sodium hypochlorite solution (50 mL, available
chlorine 5% minimum) was added dropwise with the stem of
the addition funnel extended to the bottom of the flask. When
addition was completed, the solid (3.0 mg) was collected and
identified as 6 by mass spectroscopy. Electron impact-MS
results suggested that 6 was the only component in the
recovered solid. Infrared spectra of the recovered solid and 6
were identical.
(4) Perry, R. J .; Wilson, B. D.; Miller, R. J . J . Org. Chem. 1992,
57, 2883-2887.
(5) Shen, T.-Y.; Li, J . P.; Dorn, C. P., J r. U.S. Patent 3,947,582,
March 30, 1976.
(6) Film thickness was estimated with the density of the
compounds to be 1.5 g/cm3. Im, J .; Percha, P. A.; Yeakle, D.
S. In The Materials Science and Engineering of Rigid-Rod
Polymers; Adams, W. W., Edy, R. K., McLemore, D. E., Eds.;
Symposium Proceedings, Vol. 134; Materials Research Soci-
ety: Pittsburgh, 1989; pp 307-312.
Con clu sion
Aryl benzobisoxazole and aryl benzobisthiazole com-
pounds are rigid-rod molecules that readily π-stack.
Their very low or even complete lack of stability in
common organic solvents imposes difficulties in studying
these compounds. With a combination of liquid chro-
matography, several mass spectroscopy techniques, and
the use of controls, we demonstrated that completely
different chemical reactions resulted when aryl benzo-
bisoxazole and aryl benzobisthiazole compounds were
photolyzed in the solid state, where the molecules could
stack, and in dilute solution, where π-stacking did not
happen.
(7) Yang, H. H. Kevlar Aramid Fiber; Wiley: New York, 1993;
Chapter 4.
(8) So, Y. H.; Martin, S. J .; Chau, C. C.; Wessling, R. A.; Sen, A.;
Kato, K.; Roitman, D. B.; Rochefort, W. E. U.S. Patent
5,552,221, Sept 1996. Iodide compounds, cupric compounds,
and mixtures thereof, which can undergo reversible electron
transfer, are also effective in enhancing PBO fiber photosta-
bility.
(9) Gu¨sten, H.; Rinke, M.; Kao, C.; Zhou, Y.; Wang, M.; Pan, J .
Opt. Commun. 1986, 59, 379-384.
(10) Osaheni, J . A.; J enekhe, S. A. J . Am. Chem. Soc. 1995, 117,
7389-7398.
Combining these results with our previously pub-
lished results, we conclude that intermolecular π-stack-
ing interactions between the planar molecules led to
excimer formation in the solid state. A photoinduced
electron transfer mechanism is consistent with this
dramatically different reactivity, the observed ESR
signal, and the anodic instability of these materials.
Benzo[1,2-d:5,4-d′]bisoxazoles and benzo[1,2-d:4,5-d′]-
bisthiazoles were products from heterogeneous dissocia-
tion of p-phenylenebenzobisoxazole cation radical and
p-phenylenebenzobisthiazole cation radical followed by
hydrogen abstraction. In the absence of oxygen, the ion
radical pair underwent back electron transfer, which
explains the photostability of these compounds under
nitrogen. The subsequent reactions from this photoin-
duced electron transfer reaction were retarded by
reversible redox reagents such as ferrocene com-
pounds.21
(11) (a) DePra, P. A.; Gaudiello, J . G.; Marks, T. J . Macromolecules
1988, 21, 2297-2299. (b) DePra, P. A. Ph.D. Thesis, North-
western University, J une 1989.
(12) Compounds 2 and 5 did not seem to show ESR signals.
(13) Advances in Electron-Transfer Chemistry; Mariano, P. S., Ed.;
J AI Press: Greenwich, CT, 1996; Vol. 5.
(14) A photoinduced electron-transfer mechanism was proposed
for the photodegradation of polyphenylene oxide. Pickett, J .
E. In Mechanism of Polymer Degradation and Stabilization;
Scott, G., Ed.; Elsevier: Amsterdam, 1990; pp 135-167.
(15) Page, J . A.; Wilkinson, G. J . Am. Chem. Soc. 1952, 74, 6149-
6150.
(16) Gibson, H. W.; Pochain, J . M. Macromolecules 1982, 15, 242-
247.
(17) Scurlock, R. D.; Wang, B.; Ogilby, P. R.; Sheats, J . R.; Clough,
R. L. J . Am. Chem. Soc. 1995, 117, 10194-10202.
(18) (a) Abdou, M. S. A.; Holdcroft, S. Macromolecules 1993, 26,
2954-2962. (b) Holdcroft, S. Macromolecules 1991, 24, 4834-
4838.
(19) Wasserman, H. H.; Lenz, G. R. Tetrahedron Lett. 1974, 3947-
3950.
(20) Foote, C. S.; Wexler, S. J . Am. Chem. Soc. 1964, 86, 3879-
3880.
Ack n ow led gm en t. We thank the Toyobo Co. (J a-
pan) and The Dow Chemical Co. for permission to
publish this piece of work, Drs. J ang-Hi Im and C. C.
Chau for their contributions to oxygen and PBO fiber
photostability, and Ann Birch of Editech for editing.
(21) Carbon dioxide was a gaseous product from photolysis of aryl
benzobisoxazole and aryl benzobisthiazole compounds. Only
limited examples of noncombustive or nonrespiratory conver-
sion of organic compounds into CO2 were reported. Sawyer
and co-workers reported that six CO2 molecules were gener-
ated in oxygenation of hexachlorobenzene by a superoxide
ion in aprotic media.22 In photooxidation of poly(phenylene
oxide) polymer, five molecules of oxygen were consumed per
oxidized repeating unit, and CO2 was detected as a gaseous
product.14
Refer en ces a n d Notes
(1) (a) Scheffer, J . R. Can. J . Chem. 2001, 79, 349-357. (b) Ito,
Y. Solid-Solid Organic Photochemistry of Mixed Molecular
Crystals. In Organic Molecular Photochemistry; Ramamur-
thy, V., Schanze, K. S., Eds.; Marcel Dekker: New York,
1999; Vol. 3, pp 1-70. (c) Keating, A. E.; Garcia-Garibay, M.
(22) (22) Sugimoto, H.; Matsumoto, S.; Sawyer, D. T. J . Am. Chem.
Soc. 1987, 109, 8081-8082.
MA021045I