Covalently linked acceptor - Donor systems based on isoquinoline N-oxide acceptor: Photoinduced electron transfer produces dual-channel luminescent systems that evolve chemically to photohydroxylation of the aromatic donor

Article abstract of DOI:10.1021/jo034074f

Acceptor-donor compounds containing the isoquinoline N-oxide acceptor and (methoxy)_nbenzene (n = 0, 1, 2, 3) electron donors were studied. The two chromophores are connected by a CH₂ bridging unit. All acceptor-donor compounds exhibi

Full text of DOI:10.1021/jo034074f

Cova len tly Lin k ed Accep tor -Don or System s Ba sed on
Isoqu in olin e N-oxid e Accep tor : P h otoin d u ced Electr on Tr a n sfer
P r od u ces Du a l-Ch a n n el Lu m in escen t System s th a t Evolve
Ch em ica lly to P h otoh yd r oxyla tion of th e Ar om a tic Don or
Daniel Collado, Ezequiel Perez-Inestrosa,* and Rafael Suau
Department of Organic Chemistry, Faculty of Sciences, University of Malaga, 29071-Malaga, Spain
inestrosa@uma.es
Received J anuary 22, 2003
Acceptor-donor compounds containing the isoquinoline N-oxide acceptor and (methoxy)nbenzene
(n ) 0, 1, 2, 3) electron donors were studied. The two chromophores are connected by a CH2 bridging
unit. All acceptor-donor compounds exhibit photoinduced electron transfer in acid medium that
results in the formation of a charge-transfer (CT) state. Measurements of the corresponding
electronic emission spectra revealed that these bichromophoric systems exhibit a dual fluorescence
that is strongly dependent on the protonation of the N-oxide function and the donor ability. The
CT state responsible for the red-shifted luminescence in the studied compounds is directly connected
with the initial excited state S1. On the basis of the spectroscopic and photochemical evidence,
N-O bond scission is the dominant primary photochemical process involving the CT state, the
subsequent radical coupling resulting in efficient aromatic hydroxylation. The outcome of both
quenching and sensitization experiments confirms this assertion. The results strongly suggest that
the ensuing photohydroxylation reaction is not a concerted process, but rather a two-step N-O
bond scission followed by C-O bond formation, which is regioselectively guided by the electronic
distribution of the resulting donor cation-radical.
SCHEME 1. P h otoin d u ced Ch a r ge Sh ift
In tr od u ction
Electron transfer (ET) is the most elementary and
ubiquitous of all chemical reactions, and plays a crucial
role in many essential biological processes.1 One specially
significant type of ET is the photoinduced charge shift
between two chromophores (or redox sites) of which one
is positively charged initially (Scheme 1). These processes
are classified as charge-shift electron-transfer reactions
as there is no difference in net charge between the initial
and final states.
nuclear functions in determining the factors of intercom-
ponent ET processes.4
The authors have examined a group of bichromophoric
systems in which the acceptor (A+) is an isoquinoline
N-oxide derivative, the donor (D) a benzyl derivative, and
the bridge (S) a methylene group as a model system for
an A+-S-D family of molecules such as 1a -g. The
photophysical behavior of these systems is of special
interest since coordination of the N-oxide to an electro-
phile gives rise to a CT state emission. We recently
reported the synthesis of a new type of A+-S-D podand
system5 and found it to be an efficient dual-channel
fluorosensor for Li+, Mg2+, and Ca2+, the selectivity of
the N-oxide coordination being crucial. Current work in
this context is focusing on the research for chemical
materials that can be used to develop supramolecular
photochemical systems with unique properties as optical
sensors and/or switches.6
The significance of these processes lies in the photo-
induced generation of charge-transfer states (CT), which
transforms light energy into useful chemical potentials.2
The mechanistic insights derived and molecular photo-
devices developed have fostered the study of systems in
which the acceptor and donor groups are mutually
connected via covalent linkages.3 Covalently linked ac-
ceptor-spacer-donor (A-S-D) multichromophoric sys-
tems have so far been discussed largely in mechanistic
terms, with emphasis on the roles of electronic and
The heteroaromatic N-oxide function is one of the few
with which a monomolecular photochemical reactivity is
(1) Balzani, V., Ed. Electron Transfer in Chemistry; Wiley-VCH:
Weinheim, Germany, 2001.
(2) Balzani, V.; Scandola, F., Eds. Supramolecular Photochemistry;
Ellis Horwood, Ltd.; Chischester, UK, 1991.
(4) Reference 1, Vol. 3 and references therein.
(3) (a) Closs, G. L.; Miller, J . R. Science 1988, 240, 440. (b) J ordan,
K. D.; Paddon-Row, M. N. Chem. Rev. 1992, 92, 395-410. (c) Wasielew-
ski, M. R. Chem. Rev. 1992, 92, 435-461. (d) Gust, D.; Moore, T. A.;
Moore, A. L. Acc. Chem. Res. 1993, 26, 198-205. (e) Paddon-Row, M.
N. Acc. Chem. Res. 1994, 27, 18-25.
(5) Collado, D.; Perez-Inestrosa, E.; Suau, R.; Desvergne, J .-P.;
Bouas-Laurent, H. Org. Lett. 2002, 4, 855-858.
(6) Ramamurthy, V.; Schanze, K. S., Eds. Molecular and Supramo-
lecular Photochemistry. Vol. 7. Optical Sensors and Switches; Marcel
Deckker, Inc.; New York, 2001.
10.1021/jo034074f CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/01/2003
3574
J . Org. Chem. 2003, 68, 3574-3584

Covalently Linked Acceptor-Donor Systems
CHART 1
SCHEME 2. Syn th etic Meth od ology for th e
N-Oxid es 1a -f
associated.7 Reported photochemical reactions fall into
two groups, i.e., rearrangement and N-deoxygenation.8
There is now general agreement that the former involves
the first excited singlet state and the latter probably the
triplet state.7,9 Oxygen transfer to various substrates may
be an efficient intermolecular process for diazines10 and
polyazapolycyclic N-oxides.11 Intramolecular oxygen trans-
fer from pyridine N-oxides to bridged aromatic nuclei also
has been reported.12 More recently, we reported for the
first time this type of photoreactivity in the bridged
isoquinoline N-oxide derivative Papaverine N-oxide and
showed the formation of an emissive CT state after the
protonation of the N-oxide moiety, with ensuing intramo-
lecular hydroxylation in high isolated yields.13
In this work, we examined the photophysics and
photochemistry of the benzyl derivatives of the isoquino-
line N-oxides 1a -f in terms of their luminescent nature
and of the photoreaction products obtained. Emissive
intermediate CT states were characterized and revealed
that the dual-channel emissions can be modulated simply
by modulating the redox properties of the acceptor and
donor chromophores. The efficiency of photoinduced
aromatic hydroxylation was determined and found to be
unequivocally consistent with a nonconcerted hydroxyl
radical mechanism.
Resu lts
Various mechanisms have been proposed for the pro-
cess, including the formation of a labile oxaziridine
intermediate behaving as a strong oxidizer,14 the libera-
tion of oxygen in an active form (“oxene”),15 a radical
pathway,16 and an electron-transfer pathway.11,13 The
results support the respective proposed mechanism, but
some data can always inevitably be justified in terms of
alternative mechanisms. Consequently, a number of
questions about the reaction remain open. The impor-
tance of this reaction lies in the fact that the photolysis
of aromatic N-oxides is used as a model for biological
oxidation with monooxygenases.17 Specifically, the hy-
droxylation of aromatics is assumed to take place via an
arene oxide intermediate, as shown by using deuterated
substrates, and the NIH shift mechanism is observed14a,15a
as in microsomal oxidations.18
The 1-benzylisoquinoline N-oxide derivatives 1a -f
were synthesized as shown in Scheme 2. By following the
reported procedure,19 the 1-benzylisoquinolines 2a -f
were obtained in a 70% global yield by condensation of
the isoquinoline Reissert with an appropriate benzyl
chloride, followed by hydrolysis. Oxidation of 2a -f to the
corresponding N-oxides 1a -f was accomplished with
MCPBA7 in 80-90% yields.
Stea d y-Sta te Absor p tion Sp ectr oscop y. The elec-
tronic absorption spectra for compounds 1a -f were
recorded from CH2Cl2 solutions and found to be very
similar. The corresponding wavelength maxima and
molar absorption coefficients are listed in Table 1. Figure
1 shows the electronic spectra for 1a as a typical example;
this compound exhibited a longest wavelength absorption
maximum at 366 nm. In this work, we were primarily
interested in this longest wavelength absorption since
this weak band, which extends to 380 nm, is the N-oxide
group π,π* transition, of 1Lb symmetry and with a strong
internal CT character20 (a transition from a π orbital that
is highly localized on the oxygen atom to one delocalized
on the ring).21
The absorption spectra for the N-oxides 1a -g are quite
consistent with the combined spectra for the two-bridged
units. Thus, the spectra for 1d consist of the overlapping
absorptions of isoquinoline N-oxide and p-methylanisole,
except for a small 2-nm red shift in the former (Figure
2), which suggests that the covalent linkage leads to a
slight electronic interaction in the ground state (weak
electronic coupling). Thus, direct excitation of compounds
1a -g with λ > 300 nm causes an initial local excitation
of the isoquinoline N-oxide chromophore.
(7) Albini, A.; Pietra, S., Eds. Heterocyclic N-oxides; CRC Press: Boca
Raton, FL, 1991.
(8) (a) Spence, G. G.; Taylor, E. C.; Buchart, O. Chem. Rev. 1970,
70, 231-265. (b) Albini, A.; Alpegiani, M. Chem. Rev. 1984, 84, 43-
71. (c) Bellamy, F.; Streith, J . Heterocycles 1976, 4, 1391-1447.
(9) Tokumura, K.; Matsushita, Y. J . Photochem. Photobiol. A: Chem.
2001, 140, 27-32.
(10) Tshuchiya, T.; Ari, H.; Igeta, H. Tetrahedron Lett. 1970, 2213-
2216.
(11) (a) Sako, M.; Shimada, K.; Hirata, K.; Maki, Y. Tetrahedron
Lett. 1985, 6493-6496. (b) Sako, M.; Shimada, K.; Hirata, K.; Maki,
Y. J . Am. Chem. Soc. 1986, 108, 6039-6041. (c) Sako, M.; Shimada,
K.; Hirata, K.; Maki, Y. Tetrahedron Lett. 1986, 3877-3880.
(12) Sammes, P. G.; Serra-Errante, G.; Tinker, A. C. J . Chem. Soc.,
Perkin Trans. 1 1978, 853-863.
(13) Souto-Bachiller, F. A.; Perez-Inestrosa, E.; Suau, R.; Rico-
Gomez, R.; Rodriguez-Rodriguez, L. A.; Coronado-Perez, M. E. Photo-
chem. Photobiol. 1999, 70, 875-881.
(14) (a) Akhtar, N. M.; Boyd, D. R.; Neill, J . D.; J erina, D. M. J .
Chem. Soc., Perkin Trans.
1 1980, 1693-1699. (b) Kaneko, C.;
Yamamori, M.; Yamamoto, A.; Hayashi, R. Tetrahedron Lett. 1978,
2799-2802.
(15) (a) Ogawa, Y.; Iwasaki, S.; Okuda, S. Tetrahedron Lett. 1981,
2277-2280. (b) Bucher, G.; Scaiano, J . C. J . Phys. Chem. 1994, 98,
12471-12473.
Effects of P r oton a tion . It has been reported that the
π,π* transition band at 366 nm for the N-oxide group
(16) Lin, S. K.; Wang, H. Q. Heterocycles 1986, 24, 659-664.
(17) (a) Streith, J .; Danner, B.; Sigwalt, C. Chem. Commun. 1967,
979-980. (b) Igeta, H.; Tsuchiya, T.; Yamada, M.; Arai, H. Chem.
Pharm. Bull. 1968, 16, 767-769. (c) J erina, D. M.; Boyd, D. R.; Daly,
J . W. Tetrahedron Lett. 1970, 457-460.
(18) Daly, J . W.; J erina, D. M.; Witkop, B. Experientia 1972, 28,
1129-1149.
(19) Shamma, M. The Isoquinoline Alkaloids. Chemistry and Phar-
macology; Academic Press: New York, 1972.
(20) Maier, J . P.; Muller, J . F.; Kubota, T.; Yamakawa, M. Helv.
Chim. Acta 1975, 58, 1641-1648.
(21) (a) Brand, J . C. D.; Tang, K. T. J . Mol. Spectrosc. 1971, 39,
171-174. (b) Bist, H. D.; Parihar, J . S.; Brand, J . C. D. J . Mol.
Spectrosc. 1976, 59, 435.
J . Org. Chem, Vol. 68, No. 9, 2003 3575

Collado et al.
TABLE 1. Absor p tion a n d F lu or escen ce Ma xim a a n d Qu a n tu m Yield s for 1a -f
1a
1b
1c
1d
1e
1f
a
λabs (ꢀ)b
neutral
acidic
neutral
368 (8.9)
336 (41.9)
366 (8.4)
336 (51.8)
366 (8.7)
336 (45.6)
368 (9.5)
336 (46.5)
366 (9.5)
336 (44.7)
368 (7.9)
336 (37.0)
λflc (φ)d
399 (4)
504 (5)
λ
exc 330 nm
396 (10)
394 (5)
396 (4)
394 (8)
393 (5)
acidic
380 (2)
512 (3)
512 (72)
λ
λ
exc 330 nm
exc 366 nm
384 (20)
449 (18)
380 (1)
380 (1)
380 (2)
382 (4)
452 (30)
451 (20)
479 (20)
500 (54)
a
b
d
Absorption maximum, nm. ꢀ % 10-2, M-1 cm-1
.
c Emission maximum, nm. φ × 103.
FIGURE 3. Fluorescence spectra for compounds 1a-e. Curves
F IGURE 1. Electronic absorption spectra for 1a in CH2Cl2
(solid line) and CH2Cl2 + TFA (dotted line). The inset shows
the 310-400-nm region for the 1a , 1d , 1e, and 1f N-oxides in
CH2Cl2 + TFA solutions, normalized at 336 nm.
have been scaled for clarity.
shift in this long-wavelength tail band, which is assigned
to an intramolecular CT transition. However, there was
substantial overlap between the relatively weak CT
absorption and the strong end-absorption of the proto-
nated acceptor moiety.
Stea d y-Sta te F lu or escen ce Sp ectr oscop y. In con-
trast to the absorption spectra the fluorescence spectra
for the 1a -g series exhibit differential features. Thus,
the fluorescence spectra measured in CH2Cl2 for the five
first compounds 1a -e containing the isoquinoline N-
oxide chromophore as acceptor, and the benzene ring as
the donor with zero (1a ), one (1b-d ), and two (1e)
methoxy groups as susbtituents, exhibit a nonstructured
emission band at 394 nm (Figure 3). This is the locally
excited (LE) emission of the common isoquinoline N-oxide
chromophore. It should be noted that regardless of the
excitation wavelength, the same superimposable LE
emission bands are observed for all compounds. This
indicates that excitation at short wavelength results in
energy transfer between the excited state of the donor
chromophore and the ground state of the isoquinoline
N-oxide chromophore.
In the presence of 0.1 M TFA, the emission behavior
is dependent on the excitation wavelength (Figure 4). The
LE emission of the protonated isoquinoline N-oxide is
blue-shifted to 382 nm at λexc ) 330 nm and is completely
displaced by a new charge transfer (CT) band beyond 400
nm at λexc > 360 nm (Table 1).
The fluorescence spectrum for 1f in CH2Cl2 (Figure 5)
exhibits both the LE (397 nm) and the CT (504 nm)
emission profile. This latter emission is clearly enhanced
by the presence of TFA. A double fluorescence emission
with a decreased LE emission yield (blue shifted to 380
nm) in the protonated isoquinoline N-oxide chromophore
F IGURE 2. Absorption spectra for 1d , compared with their
chromophoric subunities. Absorption spectrum for a 10-5
M
CH2Cl2 solution of 1d . (a, b) Absorption spectra for isoquinoline
N-oxide (acceptor subunit, A) and p-methylanisole (donor
subunit, D) recorded from 10-5 M CH2Cl2 solutions. (c) Absorp-
tion spectrum for an equimolar mixture of isoquinoline N-oxide
and p-methylanisole.
undergoes a strong blue shift in polar solvents,7 and the
spectrum in acids (0.1 M TFA in CH2Cl2) is blue shifted
to 336 nm (Figure 1). In addition, a weak absorption tail
appears in the 340-390-nm region on the long-wave-
length side of the protonated acceptor absorption. A
comparison of the electronic absorption spectra for the
four N-oxides 1a , 1d , 1e, and 1f (inset, Figure 1) reveals
that electronic interaction between the chromophores
increases with increasing donor ability and leads to the
appearance of a new absorption band in the near-UV
region. The intensity of this band is independent of the
concentration. Also, lowering the donor ionization poten-
tial (from 1a to 1f) results in a perceived bathochromic
3576 J . Org. Chem., Vol. 68, No. 9, 2003

Covalently Linked Acceptor-Donor Systems
F IGURE 4. 3D fluorescence spectra for compounds 1a ,d -f in 0.1 M TFA CH2Cl2 solutions.
SCHEME 3. P h otop r od u cts of th e N-Oxid es 1a -g
and 13C NMR spectroscopy and from mass spectra, as
well as by comparison with authentic samples in some
cases. The yields of the preparative photoreactions are
listed in Table 2. Except for 1d , the yields for the
hydroxylated compounds (3a -g) exceeded those for the
deoxygenated products (2a -g).
The photolysis of 1a gave the deoxygenated precursor
2a , with the regiosomer 3a as the sole hydroxylated
material. Their spectroscopic data and a comparison with
an authentic sample5 allowed us to unequivocally assign
their structure. Although the clear reaction mixture was
carefully analyzed, no other hydroxylated regioisomer
was detected.
The photolysis of the mono-metoxylated compounds
1b-d provided interesting results. Thus, irradiation of
the o- and m-methoxy compounds 1b,c produces two sets
of regioisomeric hydroxylated compounds (3b, 3b′ and 3c,
3c′) in addition to the corresponding deoxygenated
compounds (2b,c). Also, the photoreaction of 1d again
gave the derivative 3d as the sole hydroxylated material.
The presence of the hydroxy group farther away than the
ortho position with respect to the methylene bridge in
the aryl donor (as in structures of 3b, 3b′, and 3d ) was
a singular result. The interpretation of the spectroscopic
data allowed us to assign the position for the ring
hydroxylation. Because of the mechanistic implications,
these remarkable results had to be unambiguously veri-
F IGURE 5. Fluorescence emission spectra for 1f: (a) CH2Cl2
solution, any λexc. (b) CH2Cl2 + TFA, λexc ) 330 nm. (c) CH2Cl2
+ TFA, λexc ) 360 nm.
is observed simultaneously with a relative increase in the
yield of the CT emission at 504 nm, by excitation at 330
nm. Only the CT emission is observed at long exciting
wavelengths, however (see Figure 4).
P h otoch em istr y. Our preliminary studies on the
photochemical behavior of 1g, which was found to yield
the efficient intramolecular hydroxylation products 3g
and 3g′ (Table 2), revealed13 oxygen-free 0.1 mM CH2Cl2
solutions of 1g containing 0.1 M TFA to provide good
results. These conditions provide very clean reaction
mixtures and high yields of intramolecular hydroxylated
materials, with minimal yields in deoxygenated 2g and
quenching of the rearrangement processes.22 Irradiation
(Pyrex filter) of 1a -f under the same conditions (Scheme
3) yielded the corresponding deoxygenated 2a -g and the
parent hydroxylated 3a -g compounds. The structure of
1
the hydroxylated products 3a -g was determined by H
(22) Bremmer, J . B.; Wiriyachitra, P. Aust. J . Chem. 1973, 26, 437-
442.
J . Org. Chem, Vol. 68, No. 9, 2003 3577

Collado et al.
TABLE 2. Yield s (%)a of th e P h otop r od u cts fr om th e Ir r a d ia tion of 1a -g in Dega ssed CH2Cl2/0.1 M TF A Solu tion s
a
b
Calculated on the basis of consumed N-oxides. Overall photohydroxylation yields. c Data from ref 13.
fied, by comparison with both possible regioisomers (3d
and 4), which were synthesized as shown in Scheme 2.
A comparison of the physical and spectroscopic data
allowed us to unequivocally establish the structure of 3d
as the hydroxylated material derived from this photo-
chemical reaction.
the photochemical deoxygenation of N-oxides originates
from the triplet state. However, the hypotheses that both
reactions involve the singlet state and the same inter-
mediate (an oxaziridine), which can both transfer an
oxygen atom to the medium and rearrange further, have
often been presented. Direct experimental support, such
as that provided by sensitization or quenching experi-
ments, is scant. To gain insight into the mechanism of
the hydroxylation process, it is therefore necessary to
determine the nature of the excited states involved in
the photochemical reaction leading to the formation of
the resulting compounds.
Qu en ch in g of T1 by P ip er ylen e. The involvement
of two different excited states in the photochemical
pathway has been shown by comparing the results of the
photoreaction in the absence and presence of an iso-
quinoline N-oxide triplet quencher. The triplet excitation
energy23 for isoquinoline N-oxide was determined from
the phosphorescence maximun to be about 53.7 kcal
mol-1; by contrast, that for the protonated form13 is 61.6
kcal mol-1. In this work, we examined the photochemical
implications of the presence of piperylene (ET ) 59 kcal
mol-1),24 a commonplace triplet quencher that cannot act
as a triplet photosensitizer because of its poor absorption
characteristics and short lifetime.
On the basis of our previous results for the photochem-
istry of 1g, the photolysis of 1e gave the expected
regioisomeric mixture of hydroxylated compounds (3e
and 3e′). These compound possess the same dimethoxy-
benzene moiety as donor unit as the previously studied
1g; likewise, the two expected hydroxylated compounds,
ortho to the methylene bridge, were detected, albeit in a
more even proportion (with the former in a higher yield).
As expected, the photolysis of the trimethoxylated
compound 1f gave the only possible photohydroxylation
product 3f. In fact, 3f contains a highly functionalized
aromatic ring (four consecutive oxygenated and one alkyl
susbtituent), which is consistent with the efficient hy-
droxylation despite the steric hindrance and electronic
factors on the aromatic ring.
(23) Ono, I.; Hata, N. Bull. Chem. Soc. J pn. 1973, 46, 3658-3662.
(24) Turro, N. J . Modern Molecular Photochemistry; University
Science Books; Mill Valley, CA, 1991.
Qu en ch in g a n d Sen sitiza tion Exp er im en ts. As
stated above, there seems to be general agreement that
3578 J . Org. Chem., Vol. 68, No. 9, 2003

Covalently Linked Acceptor-Donor Systems
The dependence of the photoreaction pathway on the
presence of the sensitizer resulted in a dramatic differ-
ence in the outcome of the reaction. Benzophenone
inhibited the formation of the hydroxylated compound
3a and raised the yield in 2a . Hence, the only photopro-
cess observed was deoxygenation as the excited state that
gave rise to 3a could not be reached by direct triplet
sensitization. Consequently, the hydroxylation must arise
from an excited state that cannot be reached from the
isoquinoline N-oxide T1 state.
Discu ssion
The photochemical results obtained in this work (Table
2) provide unequivocal evidence that both deoxygenation
to 2a -f and hydroxylation to 3a -f are the general
pathways upon photoexcitation of acid solutions of the
N-oxides 1a -f. The most significant difference in the
photochemical consequences is the ratio of deoxygenation
(2) to hydroxylation (3) material obtained. Because all
the 1a -f N-oxides possess the same acceptor unit (the
protonated form of the isoquinoline N-oxide), the differ-
ence should be attributed mainly to the monoelectronic
oxidative ability of each donor moiety. In this respect,
the synthesis of this series of A+-S-D systems (1a -f)
has proved useful for elucidating the mechanistic impli-
cations on the photophysical and photochemical behavior.
P h otop h ysics. The studied systems contain two chro-
mophores that are covalently connected but not conju-
gated; also, their absorption spectra are simply the
combination of those for the isolated chromophores
(Figure 2). Obviously, these compounds experience no
extensive electronic interaction between the constituting
chromophores. Also, it should be noted that the intramo-
lecular acceptor/donor separation is not too large, and
that the 2-nm batochromic shift observed with respect
to the individual elements in the A+-S-D systems
indicates further, but insubstantial stabilization in the
ground state. The donor chromophores incorporated into
these bichromophoric molecules exhibit negligible ab-
sorption up to 300 nm, whereas the acceptor chromophore
exhibits strong absorption in the 240-280- and 290-320-
nm regions. As a result, the absorption spectra for all
the studied compounds are largely dominated by the
contribution of the isoquinoline N-oxide moiety and the
changes due to the presence of the different studied
donors cannot be fully observed.
Clearly, the situation is rather different in regard to
the electronic emission spectra. The donor chromophores
in the A+-S-D systems are nonfluorescent, presumably
as a result of extremely rapid radiationless relaxation
under concomitant energy transfer to the emissive S1
state of the acceptor. Under neutral conditions, the LE
emission due to the N-oxide chromophore prevails (Figure
3 is quite consistent with this situation for compounds
1a -e, for which only the nonstructured band at 394 nm
is obtained, whichever the excitation wavelength chosen).
As derived from the Weller equation25 for isoquinoline
N-oxide systems,13 there is no driving force for the
electron transfer in 1a -e; also, no other fluorescent
excited state can be reached by these bichromophoric
systems. In comparing the data, attention should be given
F IGURE 6. Variation of the [3]/[2] ratio with the piperylene
concentration.
TABLE 3. Ra tio of P h otoh yd r oxyla tion to
P h otod eoxygen a tion ([3]/[2]) in th e P h otolysis of 1a ,d ,f a t
Differ en t P ip er ylen e Con cen tr a tion s
[piperylene]
0
0.01
0.1
1a
1d
1f
3.6
0.2
1.2
4.1
0.4
1.2
12.5
0.6
1.9
TABLE 4. P h otod eoxygen a tion (2a ) a n d
P h otoh yd r oxyla tion (3a ) Yield s in th e P h otolysis of 1a in
th e P r esen ce of Ben zop h en on e a s Sen sitizer
photoproducts
[benzophenone]
2a
3a
0.0
0.1
16
80
57
0
For simplicity in the analysis of the reaction mixture,
we only examined the photolysis of 1a , 1d , and 1f since
they yielded only two photoreaction products, namely the
corresponding photodeoxygenated derivatives (2a , 2d ,
and 2f), and a single photohydroxylation derivative (3a ,
3d , and 3f, respectively). Furthermore, these three
compounds span the range of redox properties of the
studied compounds (1a -f). Table 3 and Figure 6 show
the yields obtained in the quenching experiments. With
1a , the 3a /2a ratio in the absence of quencher was 3.6.
The photodeoxygenation leading to the formation of 2a
was effectively quenched by the presence of piperylene,
and the photohydroxylation increased by a factor of
almost 4 by increasing the quencher concentration to 0.1
M. This was reflected in an increase in the 3a /2a ratio
to 12.5. With 1d and 1f as the substrates, quenching was
less marked, as reflected in the decreased slopes of the
plots (Figure 6). These results clearly suggest that
isoquinoline N-oxide photodeoxygenation is efficiently
quenched by piperylene and that it originates from the
triplet-excited state.
Tr ip let P h otosen sitiza tion . A sensitization experi-
ment was conduced to clarify the origin of the deoxygen-
ation and hydroxylation. 1a photolyzates were analyzed
in the presence and absence of benzophenone (ET ) 69
kcal mol-1).24 The results obtained are summarized in
Table 4.
(25) Rehm D.; Weller, A. Isr. J . Chem. 1970, 8, 259-271.
J . Org. Chem, Vol. 68, No. 9, 2003 3579

Collado et al.
SCHEME 4. P r op osed P h otoch em ica l Rea ctivity
for Heter ocyclic Ar om a tic N-Oxid es
to the effect of the medium as the observed reduction
potentials with N-oxides are pH-dependent, so their
reduction is easier under acid conditions.7 Thus, on
protonation of the oxygen atom of the N-oxide function,
reduction by one-electron transfer from the donor ground
state to the acceptor excited state is indeed possible and
an emissive CT state can be reached.
Compound 1f exhibits a special feature: it exhibits the
dual fluorescence emissions in both neutral and acid
solutions. In the latter, however, relative intensities are
increased with respect to the CT emission. Even in
neutral environments, the emission from the CT state
reveals that the electron transfer operates effectively in
this redox system. As can be inferred from the Rehm-
Weller equation,25 the reduction potential of the acceptor
is less favorable under neutral conditions (the ∆G factor
thus being less favorable), but this can be offset by the
decreased reduction potential of the trimethoxylated
donor.
populated upon the electron transfer. One common
feature of dual-channel fluorescent chemosensors is that
substrate binding enhances one emission channel at the
expense of the other. Our results constitute an exception
to this generalization as the combination of two signaling
mechanisms (N-oxide oxygen binding to chemical species
and induction of electron transfer) allows two fluorescent
emission bands to be obtained. Compounds 1a -e exhibit
both quenching in the LE emission of the isoquinoline
N-oxide and the simultaneous appearance of two inde-
pendent emission bands at short and long wavelengths.
From the logic viewpoint, the output can be high when
either protons or other cationic species are bound to the
N-oxide function and precise selection of the excitation
wavelength results in the independent appearance of the
two emitting channels. One specially significant feature
is the (apparent) absence of interaction of both emitting
channels.
On the other hand, compound 1f results in a photoionic
logic system in which both short- and long-wavelength
emission bands occur simultaneously at high excitation
energies, and only the long-wavelength emission bands
remain at less energetic excitations.
Although molecules 1a -f provide a rudimentary foun-
dation for future molecular photoionic devices, prelimi-
nary applications such as fluorosensors by incorporation
into new chemical guest systems5 have proved to be a
promising approach to the development of photochemical
systems and materials that can respond in a controllable
manner to chemical and/or optical stimuli.
P h ot och em ist r y. Interestingly, the studied com-
pounds exhibit efficient photochemical reactions (deoxy-
genation and aromatic hydroxylation) that yield two main
products, namely, the parent aromatic amines 2a -f and
the phenolic derivatives 3a -f, respectively (Scheme 3).
The photochemical reactions of heteroaromatic amine
N-oxides have been the subject of a number of publica-
tions7,8 from which it has been established that, upon
excitation to the S1 excited state, their photochemistry
arises, formally, from the rearrangement of oxaziridine
(A) leading to ring expansion, ring contraction, and
further oxygen migration around the heterocyclic nucleus
(Scheme 4).12 One attractive, plausible mechanistic in-
terpretation of the behavior of alkyl-substituted hetero-
cyclic N-oxides assumes that, in polar solvents, the next
step involves the formation of the zwitterions (B), fol-
lowed by a [1,2] shift of the R group to achieve the
isoquinolone (C). In nonpolar solvents, the pathway from
the oxaziridine is formulated as a [1,5] sigmatropic shift
to the epoxy structure (D), followed by a new [1,5] shift
The emissive contours of the protonated species 1a -e
in dichloromethane solutions are very strongly dependent
on the excitation wavelength. With excitation at short
wavelengths (λexc 330 nm), only the LE emission of the
protonated isoquinoline N-oxide was observed at ∼382
nm, blue shifted with respect to the neutral chromophore.
However, the appearance of a new red-shifted emission
band up to 400 nm, when using long-wavelength light,
reveals that an emissive CT state was reached. The
detectability of such CT fluorescence is further enhanced
by its significant Stokes shift, which tends to move it into
a region not overlapped by the residual LE emission of
the acceptor. The fluorescence data of Table 1 are shown
graphically in Figure 4. All the compounds from 1a to 1f
behave as dual-channel fluorescent systems. The LE
emission always remains at the same emission wave-
length, which is consistent with the fact that all possess
the same chromophoric acceptor, i.e., the protonated form
of the isoquinoline N-oxide. However, the long-wave-
length emission band shifts from 449 nm for 1a to 512
nm for 1f. This bathochromic shift is correlated with the
increased donor ability of the donor moiety in the 1a -f
series. The increase in the Stokes shift with decreasing
oxidation potential of the donor reflects the decrease in
energy of the corresponding CT states. The change in
Gibbs free energy (∆G) between the initial (S1) and final
state (CT) is a crucial parameter in all discussions
concerning electron-transfer processes.1 The comparisons
of LE and CT intensities for the emission bands in Figure
4 clearly show that the relative ease with which the CT
state can be populated correlates well with the simulta-
neous depopulation of the LE state. This reflects the fact
that the electron transfer operates as an efficient path-
way for depopulating the S1 state of the protonated form
of the isoquinoline N-oxide by decreasing the reduction
potential of the donor.
From the data in Table 1 and Figure 4 it clearly follows
that extensive or even complete quenching of the acceptor
fluorescence occurs in all the systems studied (1a -f).
Intramolecular electron transfer is assumed to provide
the additional decay channel responsible for this quench-
ing. The occurrence of intramolecular electron transfer
as a quenching mechanism is unequivocally supported
by the appearance of a long-wavelength fluorescence that
can be ascribed to radiative deactivation of the CT state
3580 J . Org. Chem., Vol. 68, No. 9, 2003

Covalently Linked Acceptor-Donor Systems
SCHEME 5. In ter m ed ia tes Con sid er ed in th e
In tr a m olecu la r Ar om a tic Hyd r oxyla tion via
Ir r a d ia tion of P yr id in e N-Oxid e Der iva tives
SCHEME 6. Con cer ted Mech a n ism P r op osed for
th e P h otoch em ica l Hyd r oxyla tion of P a p a ver in e
N-Oxid e
to the oxazepine (E). In competition with the rearrange-
ment reactions, S1 to T1 intersystem crossing as been
proposed as the pathway leading to a triplet biradical
(F). This has been claimed as the mechanism responsible
for the photochemical deoxygenation via oxygen transfer
from the N-oxide group to various substrates.
Despite these general statements, some authors have
failed to observe a potential oxaziridine intermediate.26
Experimental evidence suggests that the excited singlet
state is the last species common to the photoproducts C
and E. Whatever the mechanism may be, the rate of
isoquinolone formation is largely independent of the
solvent, so the solvent effect must be ascribed to a change
in the rate of oxazepine formation.
from the donor moiety to the singlet excited state of the
protonated form of the acceptor N-oxide, the fate of the
CT state 9, which is a diradical ion pair, is the homolytic
rupture of the N-O bond to transfer the hydroxyl radical
to the benzene ring. Proton losses upon rearomatization
yield the intramolecular hydroxylation product.
The results for the photochemistry of N-oxides 1a -f
in acid solutions allowed us to verify that the hydroxy-
lation of the aromatic donor moiety leading to 3a -f is a
common result for these compounds, which are ac-
companied by variable amounts of the parent deoxygen-
ated substances (2a -f). Interestingly, these systems can
undergo a single-electron-transfer reaction that results
in bond-cleavage/bond-construction with a regioselective
stipulation.
Kinetic studies15b of atomic oxygen in solution with
laser flash photolysis of pyridine N-oxide point to a
mechanism in which the solvent reacts with an invisible
species (atomic oxygen) to yield the observed transients.
Rapid release of 3O from the N-oxide functional group in
the lowest excited singlet state (S1) has also been
considered for pyridine N-oxide.9,15b These studies support
the formation of oxygen atom (accompanied by other
intramolecular processes, however); also, a number of
questions concerning these reactions still remain open.
Because the rearrangement product results mainly
from the oxaziridine (A), methods for suppressing this
photochemical pathway, by complexation, have been
developed. Several authors12,27 found the photochemical
behavior of heteroaromatic N-oxides to always be less
complicated, and high yields of the deoxygenated bases
to be obtained, in the presence of Lewis acids. The Lewis
acid is thought to bind the nucleophilic oxygen, thereby
inhibiting the undesirable rearrangement to oxaziridine
and its congeners. The photolysis of these complexes can
yield phenols, both in an inter- and in an intramolecular
way, via a NIH mechanism. However, on the basis of
these results (see Scheme 5), the reaction proceeds via
an oxenoid mechanism (i.e. the transfer of atomic oxygen)
that leads to either or both of the arene oxides 5 and 6;
these in turn can rearrange to the cyclohexanedienones
7 and 8. While intermediate 7 can only aromatize with
loss of deuterium, 8 can lose either a proton or a
deuterium atom and give rise to the NIH shift.
The detection of the hydroxylated compounds 3a -f
having the hydroxy function at specific positions of the
aromatic moiety, and the comparative analysis of the
different models of substitution, clearly reveals strict
requirements for the bond-construction step. This has two
major implications. First, the hydroxylation site is not
restricted to the ortho position with respect to the
methylene bridge. The two hydroxylated derivatives (3g
and 3g′) previously reported,13 as well as the new ones
(3a , 3c, 3c′, 3e, 3e′, and 3f) described herein, can be
justified in terms of an intermediate such as 9 (Scheme
6). Consequently, the N-O bond cleavage cannot exclude
a simultaneous Ar-O bond construction, in a concerted
process, via a reasonable six-member intermediate. Other
concerted intermediates have been proposed for quinoxa-
line28 and pteridine29 N-oxide derivatives. However,
because 3b, 3b′, and 3d have the hydroxylation site in a
remote position (meta or para to the methylene bridge),
they clearly exclude the concerted process and point to a
sequential bond-cleavage/bond-construction mechanism.
Second, the hydroxylation sites in the cation-radical that
derives from the donor moiety after the electron transfer
is consistent with the spin densities reported for the
model compounds, as shown in Scheme 7. Spin densities
based on Huckel calculations for some methoxy-substi-
In previous studies13 on the photochemistry of 1g, we
proposed (Scheme 6) that, after fast electron transfer
(28) Kurasawa, Y.; Takada, A.; Kim, H. S. J . Heterocycl. Chem. 1995,
32, 1085-1114.
(29) (a) Sako, M.; Shimada, K.; Hirota, K.; Maki, Y. Tetrahedron
Lett. 1985, 26, 6493-6496. (b) Sako, M.; Ohara, S.; Hirota, K.; Maki,
Y. J . Chem. Soc., Perkin Trans. 1 1990, 3339-3344.
(26) Lohse, C. J . Chem. Soc., Perkin Trans. 2 1972, 229-233.
(27) Hata, N.; Ono, I.; Kawasaki, M. Chem. Lett. 1975, 25-28.
J . Org. Chem, Vol. 68, No. 9, 2003 3581

Collado et al.
SCHEME 7. Sp in Den sities for Meth oxyben zen e
Ra d ica l Ca tion sa
SCHEME 8. Gen er a lized En er gy Level Dia gr a m
for Bich r om op h or ic A+-S-D System s
a
Arrows denote reported hydroxylation positions in the parent
compounds 1b-d .
reactive intermediate responsible of the photochemical
deoxygenation to the parent amines. With this premise
in mind, but taking into account that the two-step (bond-
cleavage/bond-construction) process leading to hydroxy-
lation can be interrupted if bond cleavage is followed by
hydroxy-radical migration and some deoxygenation ma-
terial can be formed via this disruptive pathway as a
result, the question arises as to which excited state drift
affects each photoproduct. To approach this question, we
carried out the quenching and sensitization experiments
described above.
Pyperilene was found to effectively quench the triplet
involved in the photoreaction and to reduce the produc-
tion of deoxygenated material (Table 3). This effect was
more marked on 1a but still appreciable on 1d and 1f
(Figure 6); it suggests that the triplet state of the
protonated isoquinoline N-oxide persists as that respon-
sible for the deoxygenation path. Clearly, the photosen-
sitization of 1a with benzophenone reveals that their
triplet state deoxygenates efficiently and that the hy-
droxylation material comes from another state that
cannot be reached from this one.
If we assume all protonated N-oxides 1a -e have the
same acceptor moiety, locating their triplet state at the
same energy level, and that decreasing the donor reduc-
tion potential produces CT states with a gradually
decreasing energy as can be inferred from their wave-
length emissions, then the CT state of 1a must be the
only one with a higher energy than the T1 state. There-
fore, the quenching of intersystem crossing to its T1 state
should result in a higher [3a ]/[2a ] ratio.
In summary, on the basis of the body of data reported
here, we can consider (Scheme 9) three main pathways
for the depopulation of the S1 excited state in the A+-
S-D systems studied, namely, LE fluorescence emission,
intersystem crossing to the T1 state, and electron transfer
to a CT state. All three are effective and depopulate S1
efficiently, since the described products characteristic of
their reactivity33 are not detected. The photoreaction of
1a -f yields two products only: photodeoxygenated and
photohydroxylated materials. Intersystem crossing to T1
is the main pathway to the photodeoxygenated material.
In competition, electron transfer from the donor moiety
tuted benzenes are shown in Scheme 7.30 Both charge31
and spin densities32 have been widely discussed in
relation to SN2Ar* and radical processes in the literature.
The directing influence of the electron-donor susbstitu-
ents can be rationalized in terms of the calculated values
for cation-radicals.
A comparison of Scheme 7 with the reaction data in
Table 2 reveals significant correlation. The high spin
density positions in 10 are consistent with the observed
hydroxylation selectivity. Compounds 3b and 3b′ cor-
respond to hydroxylation positions that agree with the
two high spin densities (0.28 and 0.12) in 10. Interest-
ingly, the high yield for 3b relative to 3b′ corresponds to
the high spin density position. N-oxide 1c gives both the
3c and the 3c′ phenol, which is also consistent with two
high spin density positions in cation-radical 11. The
possibility of reacting with two positions ortho to the
methylene bridge instead of with the para position may
account for the fact that no p-hydoxy-compound was
detected, consistent with the energy balance for the
solvent reorganization needed to achieve the para posi-
tion. Again, the high yield of the photoproduct 3c′ relative
to 3c exposes a relationship with the highest spin density
of 0.28 in cation-radical 11. Similarly, the hydroxylated-
only material 3d that derives from 1d has the hydoxyl
function in the meta position (i.e. the highest spin density
on cation-radical 12).
The results obtained in the photolysis of 1e and 1g,
both of which have the same substituents in the donor
moiety, can be rationalized in terms of the high spin
density of the resulting cation-radical (13) in the ortho
position relative to the others. The relatively low yield
of 3f (the system having the best donor and the ortho
position as the sole reactive site) can be rationalized in
terms of the resulting low spin density (0.02) of the
parent cation-radical (14).
E xcit ed St a t es In volved in t h e P h ot och em ica l
P r ocess. As noted earlier, the lowest triplet state for
aromatic amine N-oxides is usually regarded as the
(30) (a) Zweig, A.; Hodgson, W. G.; J ura, W. H. J . Am. Chem. Soc.
1964, 86, 4124-4129. (b) Sunberg, R. J . In Organic Photochemistry;
Padwa, A., Ed.; Marcel Deker, Inc.: New York, 1983; Vol. 6.
(31) Gilbert, A.; Baggott, J ., Eds. Essentials of Molecular Photo-
chemistry; Blackwell Scientific Publications: Oxford, UK, 1991.
(32) Said, A. H.; Mhalla, F. M.; Amatore, C.; Verpeaux, J .-N. J .
Electroanal. Chem. 1999, 464, 85-92.
(33) Bremner, J . B.; Wiryachitra, P. Aust. J . Chem. 1973, 26, 437-
442.
3582 J . Org. Chem., Vol. 68, No. 9, 2003

Covalently Linked Acceptor-Donor Systems
and 50 MHz, respectively) were recorded from CDCl3 solutions,
using the solvent residual proton signal as standard. TLC
analyses were performed on silica gel 60 F 256 plates and
column chromatography was carried out on silica gel 60 (70-
230 mesh). Melting points were obtained in open capillaries
and are uncorrected.
SCHEME 9. P r op osed Mech a n ism for th e
P h otoch em ica l Beh a vior of th e A+-S-D System s,
Illu str a ted for 1d
Sp ectr oscop ic Stu d ies. Samples for UV/Vis and emission
spectra were prepared in spectroscopic grade solvents and
adjusted to linear range response. Molar absorption coefficients
were determined by using concentrations of 10-4 or 10-5 M.
No fluorescent contaminants were detected upon excitation in
the wavelength region of experimental interest. Fluorescence
quantum yields were determined by comparison with 0.1 M
quinine sulfate in 0.05 M sulfuric acid as reference, and
corrected for the refractive index of the solvent. The samples
were made with an absorbance between 0.1 and 0.2 at the
excitation wavelength. 3D fluorescence spectra were recorded
by increasing the excitation wavelength for 1 nm.
Gen er a l P r oced u r e for Ir r a d ia tion of Sa m p les 1a -f.
Photolyses were performed under an argon atmosphere. An
immersion well photoreactor (Pyrex) equipped with a medium-
pressure mercury lamp (150 W) was used. Magnetically stirred
solutions (10-3 M) of the corresponding N-oxides in 0.1 M TFA
in CH2Cl2 were irradiated for 6 or 10 min. The photolyzates
were washed with aqueous NaHCO3 and H2O, and dried over
MgSO4. The solvent was evaporated under reduced pressure,
and the resulting material was dissolved in the minimal
amount of chloroform and subjected to column chromatography
with hexane/ethyl acetate (from 4:1 to 1:1). The respective
photodeoxygenated products obtained (2a -f) were identified
by comparison with the synthesized samples (see Supporting
Information). When necessary, mixtures of regioisomeric phe-
nols were further separated by preparative TLC with (47:3)
chloroform/methanol.
to the S1 excited state of the acceptor leads to an emissive
CT state that is responsible for the hydroxylated prod-
ucts. The CT state undergoes a homolytic N-O bond
scission to give the hydroxy radical and a cation-radical
pair. The cleavage, which may occur when the electron
transfer takes place, restores the isoquinoline nucleus
and gives the radical/cation-radical pairs as transient
intermediates. Bond formation by coupling of the hy-
droxyl radical and the arylic cation-radical is a radical
process that is regioselectively controlled by the spin
densities of the resulting donor cation-radicals. Proton
losses to restore aromaticity give the hydroxylated prod-
ucts.
1-(2-Hyd r oxyben zyl)isoqu in olin e (3a ). See ref 5.
1-(2-Meth oxy-5-h ydr oxyben zyl)isoqu in olin e (3b). Brown
oil; 1H NMR (200 MHz, CDCl3) δ 3.86 (s, 3H), 4.52 (s, 2H),
6.63 (d, 1H, J ) 2.4 Hz), 6.76 (dd, 1H, J ) 3.0, 8.5 Hz), 6.76
(d, 1H, J ) 8.5 Hz), 7.17 (d, 1H, J ) 5.5 Hz), 7.51 (dt, 1H, J )
1.2, 8.2 Hz), 7.65-7.77 (m, 2H), 7.78 (d, 1H, J ) 7.9 Hz), 8.05
(d, 1H, J ) 7.9 Hz); 13C NMR (50 MHz, CDCl3) δ 34.1, 56.3,
112.2, 114.1, 116.5, 120.3, 126.6, 126.9, 127.2, 127.6, 128.3,
130.6, 136.5, 139.5, 149.5, 151.5, 160.8; MS m/z (%) 265 (1)
Con clu sion s
The A+-S-D bichromophoric systems studied were
found to be dual-channel fluorescent compounds where
both LE and CT emissions can be modulated simply by
modifying the donor ability. J udicious design and syn-
thesis of these structures can provide attractive sensors
that combine and compete between photoluminescence
and photoinduced electron transfer. These sensors are
of special significance on account of their potential
biological applications; in fact, compounds 1a -f are
closely related to natural isoquinoline alkaloids, for which
some receptors have been reported.34
Deoxygenation to the parent amine and hydroxylation
of the aromatic donor moiety were found to be the sole
photochemical processes involved. The aromatic photo-
hydroxylation is a radical process that is regioselectively
guided by the residual spin density of the resulting
cation-radical, a process that is initiated by an intramo-
lecular electron transfer. The proposed mechanism pro-
vides a rational basis for our photochemical results and
also an approach to the interpretation of the controversial
oxygen walk process.12
[M]+, 234 (100), 84 (44); HRMS (FAB) m/z calcd.for C17H15
-
NO2 (M)+ 265.1103, found 265.1097.
1-(2-Meth oxy-3-h ydr oxyben zyl)isoqu in olin e (3b′). Brown
oil; 1H NMR (200 MHz, CDCl3) δ 3.87 (s, 3H), 4.68 (s, 2H),
6.49 (t, 1H, J ) 4.3 Hz), 6.77 (d, 2H, J ) 5.5 Hz), 7.46-7.66
(m, 3H), 7.79 (d, 1H, J ) 7.3 Hz), 8.10 (d, 1H, J ) 8.5 Hz),
8.46 (d, 1H, J ) 6.0 Hz); 13C NMR (50 MHz, CDCl3) δ 36.6,
55.8, 112.9, 116.3, 118.7, 120.3, 122.8, 125.2, 126.0, 127.6,
127.9, 130.8, 136.9, 139.9, 150.9, 152.9, 160.9; MS m/z (%), 265
(4) [M]+, 234 (100); HRMS (FAB) m/z calcd for C17H15NO2 (M)+
265.1103, found 265.1091.
1-(2-Hyd r oxy-5-m eth oxyben zyl)isoqu in olin e (3c). Pale
yellow solid; mp 139-142 °C; 1H NMR (200 MHz, CDCl3) δ
3.73 (s, 3H), 4.54 (s, 2H), 6.68 (dd, 1H, J ) 1.2, 5.8 Hz), 6.68-
6.92 (m, 2H), 7.56 (d, 1H, J ) 5.5 Hz), 7.68-7.84 (m, 3H), 8.34
(d, 1H, J ) 5.5 Hz), 8.41 (d, 1H, J ) 9 Hz); 13C NMR (50 MHz,
CDCl3) δ 36.1, 55.9, 110.5, 119.4, 120.2, 122.1, 125.5, 126.4,
127.6, 127.8, 130.8, 136.9, 139.9, 145.9, 149.4, 160.9; MS m/z
(%) 265 (100) [M]+, 248 (69); HRMS (FAB) m/z calcd for C17H15
-
NO2 (M)+ 265.1103, found 265.1106.
Exp er im en ta l Section
1-(2-Hyd r oxy-3-m eth oxyben zyl)isoqu in olin e (3c′). Pale
yellow solid; mp 129-131 °C; 1H NMR (200 MHz, CDCl3) δ
3.85 (s, 3H), 4.60 (s, 2H), 6.72-6.75 (m, 2H), 6.88 (t, 1H, J )
4.8 Hz), 7.54 (d, 1H, J ) 5.5 Hz), 7.61-7.73 (m, 2H), 7.81 (d,
1H, J ) 5.5 Hz), 8.33-8.42 (m, 2H); 13C NMR (50 MHz, CDCl3)
δ 35.3, 61.4, 114.1, 120.0, 121.5, 124.7, 124.9, 125.9, 127.3,
127.4, 130.1, 132.6, 136.5, 141.5, 144.7, 149.1, 159.9; MS m/z
Ma ter ia l a n d Equ ip m en t. All starting material and
reagents were used as received. Solvents were spectrophoto-
metric or HPLC grade and were used without further purifica-
tion. HRMS are reported as m/z. Accurate masses are reported
for the molecular ion (M + 1). H and 13C NMR spectra (200
1
(%) 265 (43) [M]+, 248 (100); HRMS (FAB) m/z calcd for C17H15
-
(34) Benavides, J .; Quateronet, D. J . Neurochem. 1983, 41, 1744-
1750.
NO2 (M)+ 265.1103, found 265.1093.
J . Org. Chem, Vol. 68, No. 9, 2003 3583

Collado et al.
1-(3-Hyd r oxy-4-m eth oxyben zyl)isoqu in olin e (3d ). Pale
yellow solid; mp 165-167 °C; 1H NMR (200 MHz, CDCl3) δ
3.79 (s, 3H), 4.56 (s, 2H), 6.72-6.78 (m, 3H), 7.46-7.53 (m,
2H), 7.62 (dt, 1H, J ) 1.2, 5.5 Hz), 7.78 (d, 1H, J ) 7.9 Hz),
8.13 (d, 1H, J ) 8.5 Hz), 8.40 (d, 1H, J ) 5.5 Hz); 13C NMR
(50 MHz, CDCl3) δ 41.0, 55.9, 109.9, 110.8, 114.9, 116.5,119.8,
120.0, 122.2, 124.8, 126.0, 127.3, 127.4, 130.2, 130.7, 132.4,
136.7, 141.1, 145.2, 145.7, 160.2; MS m/z (%) 265 (62) [M]+,
264 (100); HRMS (FAB) m/z calcd for C17H15NO2 (M)+ 265.1103,
found 265.1115.
(%) 295 (22) [M]+, 293 (27), 280 (100); HRMS (FAB) m/z calcd
for C18H17NO3 (M)+ 295.1209, found 295.1222.
1-(2-Hyd r oxy-3,4,5-tr im eth oxyben zyl)isoqu in olin e (3f).
1
Colorless oil; H NMR (200 MHz, CDCl3) δ 3.76 (s, 3H), 3.82
(s, 3H), 3.94 (s, 3H), 4.53 (s, 2H), 6.58 (s, 1H), 7.56 (d, 1H, J )
6 Hz), 7.64-7.85 (m, 3H), 8.36-8.43 (m, 2H); 13C NMR (50
MHz, CDCl3) δ 31.6, 55.8, 60.4, 122.5, 123.6, 127.2, 128.1,
128.5, 128.8, 129.1, 129.3, 132.5, 136.3, 136.4, 146.5, 152.9;
MS m/z (%) 325 (100) [M]+, 168 (72), 167 (79); HRMS (FAB)
m/z calcd for C19H19NO4 (M)+ 325.1314, found 325.1325.
Sen sitiza tion a n d Qu en ch in g Exp er im en ts. Solutions
were prepared as described in the general procedure for
irradiations and the concentration of quencher or sensitizer
was adjusted as stated in the text. Photolyzates were analyzed
by chromatography, as described above.
1-(2-Hyd r oxy-4,5-d im eth oxyben zyl)isoqu in olin e (3e).
1
Colorless oil; H NMR (200 MHz, CDCl3) δ 3.79 (s, 3H), 3.80
(s, 3H), 4.49 (s, 2H), 6.58 (S, 1H), 6.80 (s, 1H), 7.55 (d, 1H, J
) 5.5 Hz), 7.67-7.85 (m, 3H), 8.33 (d, 1H, J ) 5.5 Hz), 8.41
(d, 1H, J ) 8.3 Hz); 13C NMR (50 MHz, CDCl3) δ 36.6, 56.4,
57.6, 103.6, 114.8, 116.2, 120.8, 125.8, 125.9, 128.3, 128.5,
131.5, 137.6, 139.6, 140.4, 142.8, 149.8, 151.9, 161.9; MS m/z
(%) 295 (97) [M]+, 294 (89), 280 (100); HRMS (FAB) m/z calcd
for C18H17NO3 (M)+ 295.1209, found 295.1200.
Ack n ow led gm en t. The authors acknowledge finan-
cial support from Spain’s DGES (Project PB97-1077)
and DGI (Project BQU2001-1890).
1-(2-Hyd r oxy-3,4-d im eth oxyben zyl)isoqu in olin e (3e′).
1
Colorless oil; H NMR (200 MHz, CDCl3) δ 3.77 (s, 3H), 3.89
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails for the synthetic work and characterization data for all
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
(s, 3H), 4.55 (s, 2H), 6.36 (d, 1H, J ) 8.5 Hz), 6.94 (d, 1H, J )
8.5 Hz), 7.55 (d, 1H, J ) 6.0 Hz), 7.62-7.73 (m, 2H), 7.81 (d,
1H, 7.9 Hz), 8.33-8.41 (m, 2H); 13C NMR (50 MHz, CDCl3) δ
35.6, 55.9, 60.8, 103.3, 109.9, 118.9, 120.4, 124.2, 125.6, 126.9,
127.6, 127.9, 130.7, 131.1, 137.1, 139.4, 150.3, 160.9; MS m/z
J O034074F
3584 J . Org. Chem., Vol. 68, No. 9, 2003