An unusual photo-oxidation pathway for a poly(2,6-dimethyl-1,4-phenylene oxide) model compound

Article abstract of DOI:10.1002/pola.28608

The long wavelength UV photochemistry was investigated of a model compound for poly(2,6-dimethyl-1,4-phenylene oxide). Irradiation of the phenyl-capped dimer of 2,6-dimethyl phenol at wavelengths >290 nm with UVA-340 fluorescent lamps in the absence of oxygen gave no detectable products after 27 days. Very low conversion to oxidation products was found in the presence of oxygen, but about 20% conversion to products in which solvent had added to the benzylic methyl groups occurred under aerobic conditions when the solvents had readily abstractable hydrogen atoms. A mechanism is proposed involving the facile, reversible abstraction of a benzylic hydrogen by oxygen as a first step in the oxidation. The hydroperoxyl radical that is formed can abstract hydrogen atoms from 2° and 3° carbons of the solvent, and these radicals can combine with the benzylic radicals to give the solvent adducts.

Full text of DOI:10.1002/pola.28608

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
An Unusual Photo-Oxidation Pathway for a Poly(2,6-dimethyl-1,
4-phenylene oxide) Model Compound
James E. Pickett*
1GE Global Research, 1 Research Circle, Niskayuna, New York 12309
Correspondence to: J. E. Pickett (E-mail: dr.james.pickett@gmail.com)
Received 10 March 2017; accepted 25 March 2017; published online 00 Month 2017
DOI: 10.1002/pola.28608
ABSTRACT: The long wavelength UV photochemistry was
investigated of a model compound for poly(2,6-dimethyl-1,4-
phenylene oxide). Irradiation of the phenyl-capped dimer of
2,6-dimethyl phenol at wavelengths >290 nm with UVA-340
fluorescent lamps in the absence of oxygen gave no detectable
products after 27 days. Very low conversion to oxidation prod-
ucts was found in the presence of oxygen, but about 20% con-
version to products in which solvent had added to the benzylic
methyl groups occurred under aerobic conditions when
mechanism is proposed involving the facile, reversible abstrac-
tion of a benzylic hydrogen by oxygen as a first step in the oxi-
dation. The hydroperoxyl radical that is formed can abstract
hydrogen atoms from 28 and 38 carbons of the solvent, and
these radicals can combine with the benzylic radicals to give
C
V
the solvent adducts.
2017 Wiley Periodicals, Inc. J. Polym.
Sci., Part A: Polym. Chem. 2017, 00, 000–000
KEYWORDS: free radical; hydrogen abstraction; mechanism;
photolysis; photo-oxidation; triplet oxygen
the solvents had readily abstractable hydrogen atoms.
A
1,5-hydrogen transfer pathway shown in Scheme 2, to give an
ortho quinone methide product.11,12 Note that this product
could also arise from ether cleavage followed by abstraction of
a methyl hydrogen by the phenyl radical, also shown in Scheme
2. Quinone methides have been proposed as isolated products
in PPE photodegradation,3,4,13 but o-quinone methides gener-
ally are highly reactive14 and could be only intermediates on
the way to other products. Schneider and coworkers also
showed that PPE model compounds undergo photolysis under
both aerobic and anaerobic conditions using unfiltered Hg-Xe
lamps that included wavelengths <250 nm,12 but flash photol-
ysis studies with 266 nm excitation did not give evidence of
photolysis.15
INTRODUCTION Poly(2,6-dimethyl-1,4-phenylene oxide), com-
monly known as PPOTM resin or PPE, is an engineering ther-
moplastic usually used as blends with polymers such as high
impact polystyrene (HIPS), polyolefins, or polyamides.1 Its
thermal, thermo-oxidation, and photo-oxidation chemistry has
been extensively studied, and an accompanying paper summa-
rizes and extends the photo-oxidation work.2 Upon photo-
oxidation, PPE undergoes chain scission, crosslinking, and pro-
nounced yellowing, and several degradation mechanisms have
been proposed. Since a number of authors have suggested that
ether photolysis is an important pathway for PPE photodegra-
dation,3–5 we were interested to see whether or not a model
compound would undergo photolysis if irradiated at wave-
lengths of natural sunlight, that is, k >290 nm. As a model,
we selected the phenyl-capped dimer of 2,6-dimethylphenol
(Ph-dimer) with structure 1 in Scheme 3.
EXPERIMENTAL
UV–visible spectra were obtained on a Unicam UV-3 spectro-
photometer in 1-cm quartz cuvettes. FTIR spectra were
obtained on a Mattson Galaxy 6020 spectrometer with samples
deposited on NaCl plates from chloroform solution. Proton
NMR spectra were obtained at 300 MHz in CDCl3 solution
against a tetramethylsilane internal standard. Gas chromatog-
raphy–mass spectroscopy (GC-MS) data were obtained using a
JEOL SX 102 high-resolution, double-focusing magnetic sector
MS with a GC equipped with a 30-m DB 5 capillary column.
Aryl ether photolysis from 254 nm irradiation is well docu-
mented,6–10 and an example is shown in Scheme 1. Authors usually
have noted a substantial solvent effect on the product distributions,
but not on the types of products formed. All of the studies cited
employed 254 nm radiation in the absence of oxygen.
Schneider and coworkers proposed a different mechanism for
photolysis of aromatic ethers bearing ortho methyl groups, the
*Present Address: 4331 Buckingham Drive Schenectady, New York 12304
Additional Supporting Information may be found in the online version of this article.
C
V
2017 Wiley Periodicals, Inc.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2017, 00, 000–000
1

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
SCHEME 1 Aryl ether photolysis according to Hageman and
Huysmans.10
SCHEME 3 Preparation of Ph-dimer.
that the solution was immediately adjacent to a lamp. Sam-
ples were exposed for 27 days then analyzed by GC-MS.
Preparation of Model Compound
Phenyl-capped dimer (1, Ph-dimer; m.p. 120 8C), was pre-
pared as shown in Scheme 3 by Ullman coupling following
the procedure of Bolon.16
RESULTS
Absorbance Spectrum
Photolysis
The absorbance spectrum of the phenyl-capped dimer in
chloroform solution is shown in Figure 1. The spectrum
shows the expected cluster of B-bands in the range of 260–
295 nm with extinction coefficient of ꢀ1600 per ether link-
age. Phenyl ethers also have an E2-band (also called a K-
band) at ꢀ240 nm with extinction coefficient ꢀ10,000 and
an E1-band <190 nm with extinction coefficient >50,000.18
All these are allowed p ! p* transitions. The very weak
shoulder 310–360 nm did not disappear after repeated
recrystallizations, so it could be either a persistent impurity
or intrinsic to the compound. If real, it could be an n ! p*
transition, an absorption due to p-stacking in concentrated
solution, or an oxygen contact charge transfer complex.
Method A
Ph-dimer 1 (50 mg) was dissolved in 10 mL of isopropyl
alcohol in a 25-mL quartz tube equipped with a small mag-
netic stir bar. The solution was sparged with either N2 or O2
for 1 h then sealed with a silicone-greased ground glass
stopper secured by springs. The tubes were placed horizon-
tally on top of a UVA-340 fluorescent lamp in a Q-Lab QUVTM
cabinet for 1 month. These lamps produce a continuum
extending from 289 to >400 nm with a maximum intensity
of 0.7 Wm22nm21 at 340 nm. The intensity is comparable to
summer noon sunlight at wavelengths <340 nm, according
to Q-Lab data,17 and there is very little energy at wave-
lengths <295 nm. The temperature was about 35 8C. The
tubes were removed twice a week and agitated to ensure
good gas mixing. The reaction mixtures were analyzed by
thin-layer chromatography (TLC, silica gel; chloroform/etha-
nol eluent) and by GC-MS.
The bond dissociation energy for the ether linkage in
diphenyl ether has been reported to be 330 kJ/mol (78.8
kcal/mol).19 Fluorescence and phosphorescence data for
diphenyl ether in a glass at 77 K shows a singlet energy of
about 429 kJ/mol (102 kcal/mol) and a triplet energy of
about 338 kJ/mol (81 kcal/mol).20 Schneider et al. calculated
for the uncapped dimer a singlet energy of 424 kJ/mol (101
kcal/mol) and a triplet energy of 336 kJ/mol (81 kcal/
mol).12 If the homolysis has some activation energy, then it
begins to be questionable whether homolysis could occur
Method B
Ph-dimer 1 (25 mg) was dissolved in 5 mL of solvent in 10-
mm NMR tubes equipped with serum caps. Thin TeflonTM
tubing extending to the bottom of the tube and a needle
open to the air were inserted through the cap. Oxygen was
introduced through the tube for a short time daily to main-
tain saturation in the solvent. The tubes were mounted verti-
cally in a QUV cabinet equipped with UVA-340 lamps such
FIGURE 1 Absorbance spectrum of the model compounds in
chloroform. The molar extinction coefficients have been nor-
malized for the number of ether linkages. [Color figure can be
viewed at wileyonlinelibrary.com]
SCHEME 2 Quinone methide formation via 1,5-hydrogen trans-
fer11 or by ether cleavage.
2
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2017, 00, 000–000

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
FIGURE 2 GC-MS chromatogram (total ion current) for Ph-
FIGURE 3 Proton NMR spectrum (300 MHz, CDCl3) for Ph-
dimer photolysis in IPA/O2.
dimer photolysis in IPA/O2.
kJ/mol (4.5 kcal/mol) of each other,21 so hydrogen transfer
should be facile. (The bond dissociation energy of the alcohol
H-OR is at least 40 kJ/mol (10 kcal/mol) higher, so abstrac-
tion of the alcohol hydrogen is unlikely.) However, no prod-
ucts at all were detected. This casts considerable doubt of
ether photolysis, rearrangement, or 1,5-hydrogen transfer
occurring with irradiation wavelengths >290 nm.
from the triplet state. Irradiation from unfiltered Hg or
Xe(Hg) clearly is highly absorbed, easily populates the singlet
state, and results in efficient photochemistry. Irradiation
ꢁ254 nm actually excites into the higher energy E2 band.
Sunlight has very little energy <295 nm and is absorbed
only by the tail of the B-band.
Photolysis
An identical experiment except with oxygen in place of nitro-
gen produced an unexpected result. After exposure, TLC
showed two major product spots at lower Rf, and GC-MS
showed two major and three minor product peaks with
higher retention times than the Ph-dimer starting material
as well as a low molecular weight product determined to be
pinacol (2,3-dihydroxy-2,3-dimethyl butane) as shown in
Figure 2 and Scheme 4. The major product peaks gave parent
ions of m/z 5 376, indicating addition of IPA, with a major
fragmentation resulting from loss of m/z 5 18 (water). The
three minor product peaks gave very weak parent ions of
m/z 5 398 (2 IPA additions) with major fragments resulting
from loss of m/z 5 18 and 32 (1 and 2 molecules of water). In
addition, a very small signal was seen at retention time 8.8 min
that had m/z 5 226 corresponding to the uncapped dimer,
indicating that a small amount of ether photolysis (or cleavage
from photo-oxidation) had occurred. The mass spectra are
shown in Supporting Information Figures S1–S6.
Ph-dimer was dissolved in isopropyl alcohol (IPA), purged
with N2, sealed, and exposed to UVA-340 lamps in a QUV
cabinet under Method A. After 1 month, thin-layer chroma-
tography (TLC) and GC-MS analysis showed no change from
the starting material. Any cleavage or rearrangement prod-
ucts produced in more than 0.1% conversion should have
been easily detected by GC-MS. It was assumed that any rad-
icals formed by ether photolysis, particularly a phenyl radi-
cal, would abstract the tertiary hydrogen of IPA to produce
products rather than simply recombining. The bond dissocia-
tion energies of a hydrogen bonded to a benzyl group,
phenol, or on the central carbon of IPA are within about
The proton NMR spectra of the reaction mixture are consis-
tent with these assignments (Fig. 3). The Ph-dimer starting
material has methyl group singlets at 2.04 and 2.17 ppm.
There was no change after the photolysis under N2. Photoly-
sis under O2 in IPA gave new singlets at 2.72 and 2.55 ppm
corresponding to the benzylic methylene group of the IPA
adducts (a1 and a2 in Fig. 3). A new differentiated methyl
group singlet appeared at 2.15 ppm. Singlets corresponding
to the methyl groups adjacent to the OH group appeared at
1.26 and 1.17 ppm (b1 and b2 in Fig. 3). A broad peak at
1.59 ppm was probably the OH groups of the alcohol plus
the usual water impurity in the CDCl3. A strong singlet at
1.24 ppm can be assigned to pinacol. Integration of the ben-
zylic methylene groups indicated that about 20% of the
SCHEME 4 Addition of isopropyl alcohol to Ph-dimer upon
aerobic photolysis.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2017, 00, 000–000
3

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
TABLE 1 Photolysis of Ph-Dimer in Various Solvents Under O2
Solvent
Products
i-Propyl alcohol
2 Major mono addition,
3 minor di addition
n-Propyl alcohol
t-Butyl alcohol
Ethyl acetate
2 Major mono addition
Small amount of oxidation (by IR)
6 Mono addition
starting material had been converted to products. NMR spec-
tra of the starting Ph-dimer and the aromatic regions are
shown in Supporting Information Figures S7–S10.
FIGURE 5 Hydroxyl group region of Ph-dimer unexposed and
irradiated under O2 in several alcohol solvents. [Color figure
can be viewed at wileyonlinelibrary.com]
Photolysis under O2 was carried out using Method B and
using IPA, n-propyl alcohol, t-butyl alcohol, and ethyl acetate
as solvents as shown in Table 1. n-Propyl alcohol gave two
addition products analogous to IPA, but at apparently lower
conversion and with no detectable di-addition products. A
small amount of dimer was also detected. t-Butyl alcohol
gave no products detectable by GC-MS. IR spectroscopy, dis-
cussed below, showed some carbonyl groups arising from
oxidation, and TLC showed a trace of polar products at the
origin. Ethyl acetate gave a messy product distribution con-
taining at least six peaks with m/z 5 404 corresponding to
addition of ethyl acetate in various ways. GC-MS data are
shown in Supporting Information Figures S11–S17.
DISCUSSION
No products were found to suggest that ether photolysis is
an important pathway from irradiation with wavelengths
>290 nm. However, the addition of solvent to a benzylic
position of the model compound is a remarkable and unex-
pected reaction. For this to occur with any efficiency in the
presence of oxygen, the benzylic radical and the solvent radi-
cal probably are formed in close proximity in space and
time. This is because the concentrations of these radicals at
any time must be very low, and reaction to form the adduct
must compete with reaction with oxygen. Yet, oxygen must
be involved in some way because there is no reaction in its
absence. In addition, a mechanism must account for the very
small conversion to oxidized products with t-butyl alcohol as
the solvent yet much higher conversion to solvent adduct
products in solvents with 28 and 38 hydrogens.
FTIR spectra of the product mixtures are shown in Figures 4
and 5. The carbonyl region 1820–1680 cm21 shows the
appearance of oxidation products. The isopropyl and n-
propyl alcohol solvents, which gave addition products, had
only small amounts of oxidation products while t-butyl alco-
hol solvent gave no addition products and more oxidation.
The hydroxyl region shows a small OH band at 3600 cm21
corresponding to the alcohol of the addition products of iso-
propyl and n-propyl alcohol, but no measurable change for
products from oxidation in t-butyl alcohol.
A mechanism that seems consistent with the data is shown
in Scheme 5. We propose that absorption of a long wave-
length UV photon in the presence of oxygen results in trans-
fer of a hydrogen atom to oxygen producing benzylic and
hydroperoxyl radicals. This step may involve partial or full
charge transfer, but there is nothing to support any state-
ment on this point. In the absence of a hydrogen-donating
solvent, the most likely course is the hydroperoxyl radical
giving the hydrogen back, resulting in simply quenching of
the excited state by oxygen. Only a very small fraction of the
radical pair apparently proceeds by Path A toward oxidation
products, as evidenced by the very small conversion to oxi-
dized products seen when t-butyl alcohol is used as the
solvent.
In the presence of a hydrogen donating solvent such as IPA,
the hydroperoxyl radical can be intercepted to give hydrogen
peroxide, thus preventing the back-donation of a hydrogen
atom to the benzylic radical and diverting the reaction to
Path B. The solvent radical that is formed can combine with
the benzylic radical to form the observed addition products.
The relatively high conversion (ca. 20% in 27 days) indicates
that Path B is fairly efficient. IPA radicals that escape combi-
nation with benzyl radicals can combine to form pinacol. In
FIGURE 4 Carbonyl region of Ph-dimer unexposed and irradiat-
ed under O2 in several alcohol solvents. The spectra have been
normalized at 1601 cm21
. [Color figure can be viewed at
wileyonlinelibrary.com]
4
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2017, 00, 000–000

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
consistent with a reversible reaction with only a small fraction of
benzylic radicals going to products.
There is precedent for this type of reaction with the photo-
oxidation
of
toluene
irradiated
at
wavelengths
>285 nm.23–26 Benzaldehyde, benzyl alcohol, benzoic acid,
and 1,2-diphenylethane are formed in varying amounts
depending on the exact conditions. The kinetics suggest that
the reaction is not a free radical chain oxidation because
there is a very low activation energy, no induction period,
and strong rate dependence on oxygen concentration.24
Instead, it is suggested that the toluene-oxygen charge trans-
fer complex is excited, leading to hydrogen abstraction by
oxygen and recombination of the resulting benzylic and
hydroperoxyl radicals to form benzyl hydroperoxide, which
goes on to the other products. The diphenylethane product
is especially interesting and comprised up to 23% of the
product mix.25 Since benzyl radicals will always be in very
low concentration, it seems unlikely that two will find each
other to make such a product in fairly high conversion. This
product could more easily be formed by a mechanism like
that proposed in Scheme 5, in which toluene acts both as
the substrate and the solvent.
SCHEME 5 Proposed mechanism for methyl group oxidation.
addition, photolysis of hydrogen peroxide would generate
hydroxyl radicals that would make more IPA radicals and
more pinacol.
Adventitious radical initiation by peroxides or metals must
always be considered as an alternative mechanism. Ethyl
acetate and n-propyl alcohol are not known to form perox-
ides upon exposure to UV >290 nm, but isopropyl alcohol
has been known to oxidize when exposed to sunlight to
form hydrogen peroxide and acetone and eventually triaceto-
netriperoxide.27 Hydroxyl radicals from H2O2 decomposition
could abstract benzylic and IPA hydrogens to make radicals
that combine to make the IPA addition products. In the case
of 50 mg of Ph-dimer in 10 mL of IPA, the ratio of benzylic
hydrogens to the tertiary IPA hydrogen is 1:69. The bond
energies are nearly the same, so any radical initiator has
only a 1.4% chance of reacting with the model compound.
Most of the reaction should produce acetone and hydrogen
peroxide or pinacol. The acetone and hydrogen peroxide
would be expected to make significant amounts of triaceto-
netriperoxide. However, the GC chromatogram in Figure 2
and the NMR spectrum in Figure 3 show that the amount of
Ph-dimer-IPA adduct formed was comparable to the amount
of pinacol, and a only a trace of triacetonetriperoxide (1H
NMR singlet at 1.43 ppm)28 was detected. The concentration
of IPA free radicals cannot be very high or more of IPA oxi-
dation products would be detected. GC-MS data from other
solvents show similarly small amounts of low retention time
The reaction is quite slow by conventional standards since it
requires a month to reach significant conversion, indicating
that it occurs from excitation of a very weak absorption
band, by a formally spin-forbidden reaction, or by another
improbable process. The driving force of the reaction is the
donation of hydrogen by the solvent to the hydroperoxyl rad-
ical. In the absence of such donation (as in t-butyl alcohol),
the oxygen apparently gives back the hydrogen to the benzyl
radical. Surprisingly little oxidation product is observed, and
even less is formed in the presence of a hydrogen-donating
solvent (Fig. 4).
This mechanism is consistent with oxygen being required for
reaction, but with very little oxidation product actually
formed. It also is consistent with the benzyl and solvent rad-
icals being formed near each other in space and time since
each donates a hydrogen atom to the same molecule of oxy-
gen. Direct abstraction of a hydrogen atom by ground state
oxygen is not generally encountered and is controversial but,
in this case, we suggest it occurs from an excited state of
either the aryl ether itself or of a contact charge transfer
complex with oxygen.
The bond energy for H-O2 is 201 6 5 kJ/mol (486 1 kcal/mol),22
so the enthalpy of reaction for the first step would be 1175
kJ/mol (42 kcal/mol) from the ground states. However, the reac-
tion probably occurs through an excited state of the aryl ether.
Proceeding from the triplet state (339 kJ/mol) makes enthalpy of
reaction 2164 kJ/mol (239 kcal/mol). The reverse reaction has
a strong negative enthalpy of reaction since it proceeds through
ground states, so it should be very fast. Thus, the energetics are
FIGURE 6 Isomeric addition product not supported by the
data.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2017, 00, 000–000
5

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
compounds that could arise from combination of solvent rad-
icals (Supporting Information Figs. S11 and S15). In addition,
the fact that Ph-dimer/solvent adducts are observed in simi-
lar conversions with n-propyl alcohol and ethyl acetate sol-
vents, which are not as easily oxidized as IPA, makes this
pathway less likely.
Fellowship in 2000 that afforded the time and resources to do
this work.
REFERENCES AND NOTES
1 J. Bussink, H. T. van de Grampel, In Ullman’s Encyclopedia
of Industrial Chemistry; Wiley: New York, 1992; Vol. A21, pp
605–614.
The evidence strongly favors alcohol addition products as
shown in Scheme 4 as opposed to an isomeric structure
shown in Figure 6. This structure would not show the facile
loss of m/z 5 18 (water) seen in the MS data. The NMR
spectrum should show a singlet for the benzylic hydrogens
bearing the aliphatic ether at ꢀ4.5 ppm, where none is
found. The methyl groups of the isopropyl group would be
split into doublets near 1.2 ppm where only singlets are
observed. Thus, structures in which the alcohol has added as
an ether can be ruled out.
2 J. E. Pickett, Photo-oxidation of Poly(2,6-dimethyl-1,4-phenyl-
ene oxide): A Reinvestigation, J. Polym. Sci., Part A: Polym.
Chem. in Press.
3 F. Richter, S. Schneider, F. Distler, R. Schubert, Polym.
Degrad. Stabil. 1999, 65, 315–327.
4 A. Rivaton, Polym. Degrad. Stabil. 1995, 49, 11–20.
5 N. S. Allen, J. F. McKellar, Makromol. Chem. 1979, 180, 2875–2882.
6 R. Benn, Z. Naturforsch. 1979, 34b, 1002–1009.
7 H. J. Hageman, H. L. Louwerse, W. Mijs, J. Tetrahedron 1970,
26, 2045–2052.
8 M. S. Kharasch, G. Stampa, W. Nudenberg, Science 1952,
116, 309.
CONCLUSIONS
Photolysis of the phenyl-capped dimer of 2,6-dimethyl phe-
nol (Ph-dimer) exposed to UV wavelengths >290 nm led to
the following observations:
9 Y. Ogata, K. Takagi, I. Ishino, Tetrahedron 1970, 26, 2703–2709.
10 H. J. Hageman, W. G. B. Huysmans, Recueil 1972, 91, 528–531.
11 S. Schneider, P. Gedeck, J. Harrer, F. Richter, T. Clark,
Chem. Phys. Lett. 1996, 257, 616–621.
1. No detectable reaction in the absence of oxygen.
2. Very slight conversion to oxidation products in the pres-
ence of oxygen when in solvents with only 18 hydrogen
atoms (i.e., t-butyl alcohol).
12 S. Schneider, P. Gedeck, J. Harrer, F. Richter, T. Clark, J. Inf.
Recording 1996, 22, 521–529.
13 A. Rivaton, P. Morel, Polym. Degrad. Stabil. 1992, 35, 3–15.
14 R. W. Van de Water, T. R. R. Pettus, Tetrahedron 2002, 58,
5367–5405.
3. Less oxidation in solvents with 28 or 38 hydrogen atoms
(i.e., isopropyl alcohol, n-propyl alcohol, and ethyl acetate),
but a relatively high conversion to products with solvent
addition at the benzylic carbons.
15 S. Schneider, F. Richter, B. Brem, Polym. Degrad. Stabil.
1998, 61, 435–464.
16 D. A. Bolon, J. Org. Chem. 1967, 32, 1584–1590.
17 Q-Lab Corporation, Technical Bulletin LU-8052 (2011). http://
There is no evidence that Ph-dimer undergoes ether cleavage
when irradiated with UV wavelengths >290 nm in the absence
of oxygen for 27 days. This is consistent with work by
Schneider et al.15 in which no product transients were
observed after flash photolysis at 266 nm. In the presence of
oxygen, a very small amount of unidentified oxidation products
was formed, including a trace of uncapped dimer. Solvent
adducts to the capped dimer were formed in relatively high
conversion when a hydrogen atom donating solvent, such as
isopropyl alcohol, was used. This suggests a mechanism involv-
ing the facile, reversible abstraction of a benzylic hydrogen by
oxygen as a first step in the oxidation. The hydroperoxyl radi-
cal that is formed can abstract a hydrogen atom from 28 and 38
carbons of the solvent, and these radicals can combine with
the benzylic radicals to give the solvent adducts. Implications
for the photodegradation of poly(2,6-dimethyl-1,4-phenylene
oxide) are discussed in an accompanying paper.2
www.q-lab.com/resources/technical-bulletins.aspx;
March 20, 2017.
accessed
18 R. M. Silverstein, G. C. Bassler, T. C. Merrill, Spectrometric
Identification of Organic Compounds, 4th ed.; Wiley: New
York, 1981; pp. 305–323.
19 W. van Scheppingen, E. Dorrestijn, I. Arends, P. Mulder. J.
Phys. Chem. A. 1997, 101, 5404–5411.
20 S. K. Sarkar, A. Maiti, G. S. Kastha, Chem. Phys. Lett. 1984,
105, 355–358.
21 Handbook of Photochemistry, 3rd ed.; M. Montalti, A. Credi,
L. Prodi, M. T. Gandolfi, Eds.; Taylor & Francis: New York,
2006, pp. 530–531.
22 E. R. Fisher, P. B. Armentrout, J. Phys. Chem. 1990, 94,
4396–4398.
23 J. C. W. Chien, J. Phys. Chem. 1965, 69, 4317–4325.
24 K. S. Wei, A. H. Adelman, Tetrahedron Lett. 1969, 3297–3300.
25 L. K. Sydnes, I. C. Burkow, S. H. Hansen, Acta Chem. Scand.
1985, B 39, 829–835.
26 M. Pasternak, A. Morduchowitz, Tetrahedron Lett. 1983, 24,
4275–4278.
ACKNOWLEDGEMENTS
27 L. Bretherick, Handbook of Reactive Chemical Hazards, 4th
ed.; Butterworth-Heinemann Ltd.: London, 1990, p. 391.
The author thanks his colleagues at GE Global Research: W.
Ligon and H. Grade for obtaining the GC-MS data. He also thanks
the management of GE Global Research for a Coolidge
28 J. C. Oxley, J. L. Smith, H. Chen, Propellants Explosives
Pyrotechnics. 2002, 27, 209–216.
6
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2017, 00, 000–000