Strongly UV absorbing bifunctional azoalkanes

Article abstract of DOI:10.1021/ol016430r

(Figure Presented) Two new anthracene-containing azoalkanes (1 and 2) absorb UV light 600 times more strongly than simple azoalkanes. Intramolecular energy transfer from excited singlet anthracene to the azo group is nearly complete, but despite the close proximity of the two chromophores, 1 and 2 continue to exhibit anthracene fluorescence. Thermolysis of these compounds in the presence of monomers affords fluorescent labeled polymers. Compounds 1 and 2 are the first azoalkanes to undergo induced decomposition in solution.

Full text of DOI:10.1021/ol016430r

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ORGANIC
LETTERS
2001
Vol. 3, No. 20
3145-3148
Strongly UV Absorbing Bifunctional
Azoalkanes
Paul S. Engel,* Huifang Wu, and William B. Smith†
Departments of Chemistry, Rice UniVersity, P.O. Box 1892, Houston, Texas 77251,
and Texas Christian UniVersity, Fort Worth, Texas 76129
engel@rice.edu
Received July 13, 2001
ABSTRACT
Two new anthracene-containing azoalkanes (1 and 2) absorb UV light 600 times more strongly than simple azoalkanes. Intramolecular energy
transfer from excited singlet anthracene to the azo group is nearly complete, but despite the close proximity of the two chromophores, 1 and
2 continue to exhibit anthracene fluorescence. Thermolysis of these compounds in the presence of monomers affords fluorescent labeled
polymers. Compounds 1 and 2 are the first azoalkanes to undergo induced decomposition in solution.
Azoalkanes are widely used thermal free radical initiators,1,2
yet few applications take advantage of their photochemical
lability because of their low extinction coefficients in the
near UV.3 We reasoned that covalent attachment of an
azoalkane moiety to a strongly absorbing chromophore might
lead to a useful initiator if energy transfer from the
chromophore to the azo group were even moderately
efficient. Anthracene (AN) was selected as the donor because
its extinction coefficient in the near UV is >600 times greater
than that of azo-tert-butane (ATB). Although anthracene
inhibits the polymerization of vinyl monomers, the fact that
highly purified 1- and 2-vinylanthracenes can be polymer-
ized4 suggested that AN-containing azoalkanes would serve
as initiators. We describe herein the radical chemistry of
unsymmetrical azo compounds 1 and 2 and show that these
“chromophore-labeled initiators”5 are capable of introducing
a strongly fluorescent AN group into polystyrene (PS) and
poly(methyl methacrylate) (PMMA). Since the AN is linked
to the polymer by C-C bonds alone, it will not be removed
by hydrolysis, unlike ester-linked fluorescent polymers.5,6
Fluorescent probes are useful for investigating the physical
properties of polymers7 and energy transfer in polymers,8
and AN derivatives have been employed in the fields of
molecular recognition and sensing,9 in solid-phase synthe-
sis,10 and in studies of polymer chain dynamics.11-15
Besides examining the potential of 1 and 2 to serve as
radical initiators, we wanted to study the efficiency of
intramolecular energy transfer, which is a topic of long-
standing interest to organic photochemists.16 Our earlier
observation that triplet energy transfer was often surprisingly
slow in bichromorphic azoalkanes17 piqued our curiosity
(6) Vyprachticky, D.; Pokorna, V.; Pecka, J.; Mikes, F. Macromolecules
1997, 30, 7821.
(7) Jager, W. F.; Sarker, A. M.; Neckers, D. C. Macromolecules 1999,
32, 8791.
(8) Martin, T. J.; Webber, S. E. Macromolecules 1995, 28, 8845.
(9) Sun, S.; Desper, J. Tetrahedron 1998, 54, 411.
(10) Congreve, M. S.; Ladlow, M.; Marshall, P.; Parr, N.; Scicinski, J.
J.; Sheppard, T.; Vickerstaffe, E.; Carr, R. A. E. Org. Lett. 2001, 3, 507.
(11) Yokotsuka, S.; Okada, Y.; Tojo, Y.; Sasaki, T.; Yamamoto, M.
Polym. J. 1991, 23, 95.
(12) Sasaki, T.; Arisawa, H.; Yamamoto, M. Polym. J. 1991, 23, 103.
(13) Ben Cheikh Larbi, F.; Halary, J.-L.; Monnerie, L. Macromolecules
1991, 24, 867.
(14) Johnson, B. S.; Ediger, M. D.; Yamaguchi, Y.; Matsushita, Y.; Noda,
I. Polymer 1992, 33, 3916.
(15) Adolf, D. B.; Ediger, M. D.; Kitano, T.; Ito, K. Macromolecules
1992, 25, 867.
(16) Wagner, P. J.; Klan, P. J. Am. Chem. Soc. 1999, 121, 9626.
(17) Engel, P. S.; Horsey, D. W.; Scholz, J. N.; Karatsu, T.; Kitamura,
A. J. Phys. Chem. 1992, 96, 7524.
† Texas Christian University. Email: w.b.smith@tcu.edu.
(1) Moad, G.; Solomon, D. H. Azo and Peroxy Initiators. In Compre-
hensiVe Polymer Science; Allen, G. A., Bevington, J. C., Eds.; Pergamon
Press: Oxford, 1989; Vol. 3, Chapter 8, pp 97-146.
(2) Engel, P. S. Chem. ReV. 1980, 80, 99.
(3) Rabek, J. F. Mechanisms of Photophysical Processes and Photo-
chemical Reactions in Polymers; Wiley: New York, 1987.
(4) Stolka, M.; Yanus, J. F.; Pearson, J. M. Macromolecules 1976, 9,
710.
(5) Harth, E.; Hawker, C. J.; Fan, W.; Waymouth, R. M. Macromolecules
2001, 34, 3856.
10.1021/ol016430r CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/05/2001

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
about the analogous singlet process. Even when acyclic
azoalkanes are promoted to their excited triplet state, loss
of nitrogen is usually inefficient.18 We found many years
ago that ATB was a rapid quencher of AN fluorescence and
that irradiation of AN caused decomposition of added ATB
Table 1. Thermolysis Rates of Azo Compounds 1 and 2 with
Added Thiophenola
temp, °C
104 k, s-1
Rb
half-life, min
half-livesc
91.94d
103.09d
105.12d
111.16d
85.88e
1.16
4.17
5.16
10.0
2.58
18.3
0.99953
0.99956
0.99955
0.99847
0.99962
0.99942
99.8
27.7
22.4
11.5
44.8
6.32
2.91
3.32
3.56
3.30
3.19
5.38
with a much higher nitrogen quantum yield (ΦN )18 than the
2
triplet sensitized reaction. It therefore seemed possible that
1 and 2 would be good photochemical free radical sources.
Synthesis and Nitrogen Yields. Reaction of 2-cyanoan-
thracene19,20 with methyllithium and cerium chloride pro-
ceeded normally with no loss of the functional groups seen
in 2-cyanonaphthalene.21 Treatment of the resulting amine
with tert-butylsulfamyl chloride22 followed by hypochlorite
oxidation23 afforded 2 as a solid that melted and evolved
nitrogen sharply at 131 °C.
103.09e
a Using 29 mg of azo compound and 40 mg of PhSH in 1 mL of
1-methylnaphthalene solvent. b Correlation coefficient. c Number of half-
lives fitted to first-order line. d Compound 1. e Compound 2.
4. Reverse-phase HPLC analysis of the product mixture sup-
The more hindered 1-isomer proved to be accessible by
initial conversion of 1-acetylanthracene24 to the tertiary
alcohol. Treatment with hydrazoic acid yielded the desired
azide plus the dehydration product, 1-isopropenylanthracene.
Lithium aluminum hydride reduction of the azide afforded
the amine, which was transformed to the unsymmetrical azo
compound 1 by a similar method as 2.
ported the GC/MS results, showing a complex mixture con-
Irradiation of 1 and 2 at 313 nm in benzene afforded N2
in 85-90% yield, whereas thermolysis at 105-110 °C gave
89-93% nitrogen. The fact that these yields fall short of
100% is attributed to decomposition by ambient light prior
to sealing the tubes.
taining several compounds with the characteristic anthracene
UV spectrum. A similar product distribution was obtained
upon irradiation of 1 at 25 °C, but the amount of dispro-
portionation products was even lower than in the thermolysis.
Despite the temptation to attribute these M + 1 ) 277
products to simple radical recombination, the curved kinetic
plots suggested another mechanism: induced decomposition
of 1 by tert-butyl radicals. We tested this idea using tert-
heptyl perpivalate as a tert-butyl radical source.27,28 In
separate control experiments, thermolysis of of 1 in C6D6 at
63 °C showed no decomposition over the course of 8 h, but
under the same conditions, the perester decomposed com-
pletely in 6 h. A mixture of 1 with a 4.6-fold molar excess
of tert-heptyl perpivalate in C6D6 was likewise heated at 63
°C, leading to complete disappearance of 1 in 4 h, judging
by NMR. The largest GC peak (retention time 16.69 min)
corresponded to a previously minor product, showed M + 1
) 277 in the CI/GC/MS, and degraded to a complex mixture
when subjected to TsOH. The 16.69 min GC peak was
accompanied by all of the peaks previously seen in the GC
trace, plus others. On the basis of its mass and the observation
that HPLC-UV analysis of the perester product mixtures
showed hardly any peaks with AN absorption, we assign
the new product to a quinoid AN structure such as 5. In
support of this notion, the 1H NMR spectrum of the
thermolyzed mixture exhibited peaks in the olefinic region
(5.75-6.26 ppm).
Thermolysis Kinetics. To evaluate the thermal stability
of 1 and 2, we determined the kinetics in our automated gas
evolution apparatus.25,26 Surprisingly, the “first-order” plots
for both isomers were curved, falling more slowly as the
runs progressed. Because the curvature persisted in several
solvents and even in the presence of 1,4-cyclohexadiene, we
ran the thermolysis of 1 with added thiophenol, whereupon
the first order plots finally became linear (cf. Table 1). Fitting
the rate constants for 1 to the Eyring equation gave ∆H‡ )
30.5 kcal/mol and ∆S‡ ) 6.5 eu, whereas 2 gave ∆H‡ )
29.8 kcal/mol and ∆S‡ ) 7.8 eu. For comparison of the two
isomers, the more meaningful parameter is ∆G‡ (100 °C),
which was 28.0 kcal/mol for 1 and 26.9 kcal/mol for 2.
Thermolysis and Photolysis Products. Thermolysis of
1 in 1-methylnaphthalene with thiophenol at ∼96 °C afforded
mainly 1-isopropylanthracene. When thiophenol was omitted,
the thermolysis mixture became quite complex, revealing
minor amounts of 1-isopropylanthracene and 1-isopropenyl-
anthracene plus larger quantities of three compounds whose
CI/GC/MS (M + 1 ) 277) corresponded to recombination
of tert-butyl radicals with 3. We assign the last GC peak to the
expected R-recombination product, 1-tert-heptylanthracene
(18) Engel, P. S.; Bartlett, P. D. J. Am. Chem. Soc. 1970, 92, 5883.
(19) Figeys, H. P.; Dralants, A. Tetrahedron 1972, 28, 3031.
(20) Ullmann, F.; Sone, M. Justus Liebigs Ann. Chem. 1911, 380, 336.
(21) Ciganek, E. J. Org. Chem. 1992, 57, 4521.
(22) Hendrickson, J. B.; Joffee, I. J. Am. Chem. Soc. 1973, 95, 4083.
(23) Ohme, R.; Preuschhof, H.; Heyne, H.-U. Organic Syntheses;
Wiley: New York, 1988; Collect. Vol. VI, pp 78-81.
(24) Ni, Y.; Siegel, J. S.; Hsu, V. L.; Kearns, D. R. J. Phys. Chem. 1991,
95, 9211.
(25) Engel, P. S.; Dalton, A. I.; Shen, L. J. Org. Chem. 1974, 39, 384.
(26) Billera, C. B.; Dunn, T. B.; Barry, D. A.; Engel, P. S. J. Org. Chem.
1998, 63, 9763.
Irradiation of 2 at room temperature in C6D6 without
thiophenol and thermolysis of 2 in C6D6 at 110 °C showed
(27) Nakamura, T.; Busfield, W. K.; Jenkins, I. D.; Rizzardo, E.; Thang,
S. H.; Suyama, S. J. Org. Chem. 2000, 65, 16.
(28) Nakamura, T.; Watanabe, Y.; Tezuka, H.; Busfield, W. K.; Jenkins,
I. D.; Rizzardo, E.; Thang, S. H.; Suyama, S. Chem. Lett. 1997, 1093.
3146
Org. Lett., Vol. 3, No. 20, 2001

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
four products by GC, the first two of which were 2-iso-
propylanthracene and 2-isopropenylanthracene. Next came
two compounds with identical EI mass spectra (m/e ) 276),
corresponding to recombination of the 2-anthracenyl-2-propyl
radical 6 with a tert-butyl radical. On the basis of product
isolation and NMR spectroscopy, we assign the last GC peak
to 2-tert-heptylanthracene 7.
Table 2. Fluorescence and Absorbance for AN, 1, 2, and
Functionalized Polymers
UV
corrected
intensityb
compound
solvent intensitya
absorbance
anthracenec,d
1c,d
hexane
hexane
hexane
toluene
toluene
toluene
toluene
(1.0)
0.032
0.16
(1.0)
0.21
0.61
(1.0)
0.044
0.12e
(1.0)
0.082
0.14
0.043e
2c,d
0.14e
anthracenef,g
PS from 1f,h
PS from 2f,i
anthracenef,j
0.21 (360 nm)
0.032 (366 nm)
0.12 (360 nm)
0.46 (360 nm)
0.031 (366 nm)
PMMA from 1f,k toluene
a Using 325 nm excitation; emission spectrum integrated from 350 to
550 nm. b Corrected for varying absorbance at 325 nm. c Degassed and
sealed. d At 6.29 × 10-5 M. e At 325 nm. f Nitrogen purged for 1 min.
g Reference solution for PS fluorescence. h At .02 mg/10 mL. i At 5.11 mg/
10 mL. j Reference solution for PMMA fluorescence. k At 5.33 mg/10 mL.
Absorption and Fluorescence Spectroscopy. The excita-
tion and fluorescence spectra of 2 and the absorption spec-
trum of ATB are shown in Figure 1. The absorption maxima
325 nm where ATB has ꢀ of only 4, we found a linear Stern-
Volmer plot with kq ) 1.5 × 1010 M-1 s-1.
Polymerization Experiments. Photolysis of 9-anthryl
ketones in the presence of methyl methacrylate yielded AN-
tagged PMMA but in low chemical and quantum yield.30 In
contrast, our system produces anthracenyl-2-propyl radicals
efficiently upon either photolysis or thermolysis. Compounds
1 and 2 were used to initiate the polymerization of styrene
in benzene, yielding fluorescent polystyrene (cf. Table 2).
GPC of the PS from 1 gave Mn )26300, Mw ) 41270, while
the PS from 2 showed Mn ) 15240, Mw ) 28600. Compound
1 also served to polymerize methyl methacrylate, and like
PS, the purified polymer exhibited typical AN absorption
and emission spectra. The molecular weight Mn of the PS
found by GPC predicts that the AN absorbance should be
3-6 times higher than reported in Table 2, under the rea-
sonable assumption that the extinction coefficient of the AN
moiety is unchanged in the polymer. A possible interpretation
of this result is that a sizable fraction of the polymer contains
quinoid AN end groups, similar in structure to 5.
Figure 1. Excitation spectrum of the 398-nm emission of 2,
emission spectrum of 2, and absorption spectrum of azo-tert-butane.
Spectral intensities are adjusted to facilitate comparison.
The observation that intramolecular singlet energy transfer
is incomplete in 1 and 2 is the same as found for intramolec-
ular triplet sensitization of the azo group.17 However, this
result is even more surprising in 1 and 2 because intermo-
lecular singlet energy transfer is very fast and because the
chromophores are separated by only two bonds, which is
the shortest possible link. The small size of the acceptor chro-
mophore could be responsible for the continued fluorescence
of 1 and 2.17 The nitrogen quantum yield of 2 (0.74) was
higher than the usual 0.5 maximum for acyclic azoalkanes.2
Since singlet energy transfer is about 90% complete in 1 and
2, one would expect a maximum quantum yield of ∼0.45.
The enhanced value found experimentally can be attributed
to induced decomposition by tert-butyl radicals. The high
quantum yield for intramolecular singlet sensitized photolysis
of 2 stands in sharp contrast to triplet sensitized irradiation
of ATB, whose nitrogen quantum yield is below 0.02.18
Turning to the thermolysis results, we note that ∆G‡ (100
°C) of 1 (28.0 kcal/mol) and 2 (26.9 kcal/mol) are not much
of 2 coincide with its excitation maxima (376, 356, 340, 324
nm) and with those of AN. However, the emission maxima
are slightly red-shifted relative to those of AN. UV absor-
bances and fluorescence intensities are shown in Table 2.
On the basis of the integrated area of the emission, the
fluorescence efficiency of 1 and 2 is 12 and 7 times less
than that of AN, respectively, implying that singlet energy
was mostly transferred from AN to the azo moiety. We
therefore determined the nitrogen quantum yield of 2,
anticipating that singlet energy transfer would lead to
deazatation. Using a benzophenone-DBH actinometer at 313
nm,29 we obtained Φ ) 0.74.
ATB is a very rapid bimolecular quencher of AN
fluorescence. Although the quenching rate constant was
determined years ago18 as 2.0 × 1010 M-1 s-1, we repeated
this experiment using more dilute AN solutions in toluene
and more concentrated solutions of ATB. With excitation at
(29) Engel, P. S. J. Am. Chem. Soc. 1969, 91, 6903.
Org. Lett., Vol. 3, No. 20, 2001
(30) Modi, P. J.; Guillet, J. E. Macromolecules 1994, 27, 3319.
3147

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
lower than that of tert-butylazocumene (28.6 kcal/mol).31
This result is surprising because azoalkane thermolysis rates
generally reflect the stability of the incipient radicals, and
the C-H bond dissociation energy (BDE) of 1-methylnaph-
thalene and 9-methylanthracene lie 2.9 and 6.6 kcal/mol,
respectively, below that of toluene.32 To gain further insight
into this discrepancy, we carried out B3LYP calculations on
benzylic C-H bond dissociation and the results are sum-
marized in Table 3. While the calculated BDEs of all three
and 0.223 at the R, 1, and 9 positions, respectively, and those
of 3 are 0.599, 0.388, 0.358, and 0.152 at the R, 2, 4, and
10 positions, respectively. No other ring position exhibited a
spin density over 0.11. Ignoring steric hindrance, these values
predict several recombination products for 6 and 3. Such
hindrance does not affect AN itself, where the 9 and 10 posi-
tions are attacked preferentially by radicals.33,34 In addition to
the AN recombination products that we see by HPLC/UV
and GC/MS, there are others that lack the AN chromophore.
These are assumed to be quinoid compounds such as 5 derived
from recombination on the ring without rearomatization. The
complexity of the products from 1 and 2, the difficulty of
separating the various ANs and quinoid ANs, and their
sensitivity to acid rendered a complete analysis impractical.
Four lines of evidence point to induced decomposition of
1 and 2: the curved first-order plot without thiophenol, the
linear plot with thiophenol, the high nitrogen quantum yield
of 2, and the fact that thermolysis of tert-heptyl perpivalate
in the presence of 1 causes disappearance of the azo
compound and yields mostly non-AN products. To our
knowledge, there is no previous case of induced decomposi-
tion of an azoalkane, except for some early gas phase work.35
For example, the linear activation plot for thermolysis of
allylic azoalkanes36 argues against induced decomposition.
A likely reason why we see this complication in 1 and 2 is
that AN is highly prone to radical attack. Thus its methyl
affinity is about 16 times greater than that of a simple 1-olefin
and is over 1300 times larger than that of benzene.37
2-Cyanopropyl and benzyl radicals readily attack AN at the
9-position,33,34 supporting the notion that the 16.69 min GC
peak from the perester and 1 has the structure 5.
In summary, the new azo compounds 1 and 2 are strong
UV absorbers by virtue of their AN chromophore. On
irradiation, singlet electronic energy is mostly transferred
from the AN ring to the azo group; however, despite the
close proximity of these chromophores, energy transfer is
incomplete. Compounds 1 and 2 are effective photoinitiators
that yield tert-butyl radicals plus radicals 3 and 6. Ther-
molysis occurs at moderate temperatures, and the rate
constants can be rationalized by a combination of steric and
radical-stabilization arguments. If the tert-butyl radicals are
not scavenged, they attack the AN ring, leading to the first
case of solution phase induced decomposition of an azo
compound. Initiators 1 and 2 convert styrene and methyl
methacrylate to end-labeled fluorescent polymers.
Table 3. B3LYP Calculations on Cumene,
Isopropylanthracenes, and Their Radicals
energya
25 °C
3° C-H BDEb Erel kcal/mol
kcal/mol (298 °C)
species
(298 °C)c
H•
-0.498856
-350.002478
-349.372017
-657.182854
cumene
82.6
82.9
79.5
0.0
2.5
cumyl•
1-isopropylAN
1-isopropylAN•(3) -656.551876
2-isopropylAN -657.186782
2-isopropylAN•(6) -656.561305
(0.0)d
a In hartrees. b Bond dissociation energy. c Strain energy assuming that
cumene is unstrained. d Assumed same as cumene.
compounds are indeed similar, we can detect conformational
energy differences that form a consistent picture, despite
lying barely outside the error limits of the calculations.
Both cumyl radical and 6 are planar; hence, the 3.1 kcal/
mol lower BDE of 2-isopropylanthracene suggests greater
resonance stabilization of 6 than cumyl. On the other hand,
repulsion between methyl and the 9-hydrogen, which twists
the side chain of 3 out of plane by 15°, raises its energy 5.9
kcal/mol above that of 6. Similar steric repulsion in 1-iso-
propylanthracene elevates its energy 2.5 kcal/mol above that
of the 2-isomer, with the net result that the C-H BDE of
1-isopropylanthracene is 3.4 kcal/mol greater than that of
2-isopropylanthracene. In dissociation of 1 and 2, the radicals
must again differ by 5.9 kcal/mol, but the 1.1 kcal/mol lower
∆G‡ (100 °C) of 2 compared to that of 1 means that the
ground-state energy of 1 lies 4.8 kcal/mol above that of 2,
if entropy effects are ignored. Quite reasonably, the steric
repulsion that elevates 1-isopropylanthracene above its
2-isomer becomes more severe in 1 versus 2. Thus ground-
state steric repulsion narrows the ∆G‡ (100 °C) gap (1.1 kcal/
mol) between 1 and 2 more than expected on the basis of
the C-H BDE difference between 1-isopropylanthracene and
2-isopropylanthracene (3.4 kcal/mol).
Acknowledgment. We thank the National Science Foun-
dation and the Robert A. Welch Foundation for financial
support of this work.
Although the structure of only three products (isopropyl,
2-propenyl, and tert-heptyl anthacenes) was determined with
certainty, it is clear from GC/MS and HPLC/UV that there
are at least two radical recombination products from 6 and
three from 3. A clue about their possible structures can be
obtained from the spin density of radicals 6 and 3. The values
calculated by density functional theory for 6 are 0.643, 0.436,
Supporting Information Available: Experimental pro-
cedures and spectral data for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL016430R
(33) Bickel, A. F.; Kooyman, E. D. Rec. TraV. Chim. 1952, 71, 1137.
(34) Beckwith, A. L. J.; Waters, W. A. J. Chem. Soc. 1957, 1001.
(35) Riblett, E. W.; Rubin, L. C. J. Am. Chem. Soc. 1937, 59, 1537.
(36) Engel, P. S.; Bishop, D. J. J. Am. Chem. Soc. 1975, 97, 6754.
(37) Zytowski, T.; Fischer, H. J. Am. Chem. Soc. 1997, 119, 12869.
(31) Engel, P. S.; Pan, L.; Ying, Y.-M.; Alemany, L. B. J. Am. Chem.
Soc. 2001, 123, 3706.
(32) Stein, S. E.; Brown, R. L. J. Am. Chem. Soc. 1991, 113, 787.
3148
Org. Lett., Vol. 3, No. 20, 2001